CoverageMaster (CoM) is a copy number variation (CNV) calling algorithm based on depth-of-coverage maps designed to detect CNVs of any size in exome [whole exome sequencing (WES)] and genome [whole genome sequencing (WGS)] data. The core of the algorithm is the compression of sequencing coverage data in a multiscale Wavelet space and the analysis through an iterative Hidden Markov Model. CoM processes WES and WGS data at nucleotide scale resolution and accurately detects and visualizes full size range CNVs, including single or partial exon deletions and duplications. The results obtained with this approach support the possibility for coverage-based CNV callers to replace probe-based methods such as array comparative genomic hybridization and multiplex ligation-dependent probe amplification in the near future.

The lack of a reliable and easy-to-operate screening pipeline for disease-related noncoding RNA regulatory axis is a problem that needs to be solved urgently. To address this, we designed a hybrid pipeline, disease-related lncRNA-miRNA-mRNA regulatory axis prediction from multiomics (DLRAPom), to identify risk biomarkers and disease-related lncRNA-miRNA-mRNA regulatory axes by adding a novel machine learning model on the basis of conventional analysis and combining experimental validation. The pipeline consists of four parts, including selecting hub biomarkers by conventional bioinformatics analysis, discovering the most essential protein-coding biomarkers by a novel machine learning model, extracting the key lncRNA-miRNA-mRNA axis and validating experimentally. Our study is the first one to propose a new pipeline predicting the interactions between lncRNA and miRNA and mRNA by combining WGCNA and XGBoost. Compared with the methods reported previously, we developed an Optimized XGBoost model to reduce the degree of overfitting in multiomics data, thereby improving the generalization ability of the overall model for the integrated analysis of multiomics data. With applications to gestational diabetes mellitus (GDM), we predicted nine risk protein-coding biomarkers and some potential lncRNA-miRNA-mRNA regulatory axes, which all correlated with GDM. In those regulatory axes, the MALAT1/hsa-miR-144-3p/IRS1 axis was predicted to be the key axis and was identified as being associated with GDM for the first time. In short, as a flexible pipeline, DLRAPom can contribute to molecular pathogenesis research of diseases, effectively predicting potential disease-related noncoding RNA regulatory networks and providing promising candidates for functional research on disease pathogenesis.

Structural information for chemical compounds is often described by pictorial images in most scientific documents, which cannot be easily understood and manipulated by computers. This dilemma makes optical chemical structure recognition (OCSR) an essential tool for automatically mining knowledge from an enormous amount of literature. However, existing OCSR methods fall far short of our expectations for realistic requirements due to their poor recovery accuracy. In this paper, we developed a deep neural network model named ABC-Net (Atom and Bond Center Network) to predict graph structures directly. Based on the divide-and-conquer principle, we propose to model an atom or a bond as a single point in the center. In this way, we can leverage a fully convolutional neural network (CNN) to generate a series of heat-maps to identify these points and predict relevant properties, such as atom types, atom charges, bond types and other properties. Thus, the molecular structure can be recovered by assembling the detected atoms and bonds. Our approach integrates all the detection and property prediction tasks into a single fully CNN, which is scalable and capable of processing molecular images quite efficiently. Experimental results demonstrate that our method could achieve a significant improvement in recognition performance compared with publicly available tools. The proposed method could be considered as a promising solution to OCSR problems and a starting point for the acquisition of molecular information in the literature.

Artificial mutagenesis and protein engineering have laid the foundation for antigenic characterization and universal vaccine design for influenza viruses. However, many methods used in this process require manual sequence editing and protein expression, limiting their efficiency and utility in high-throughput applications. More streamlined in silico tools allowing researchers to properly analyze and visualize influenza viral protein sequences with accurate nomenclature are necessary to improve antigen design and productivity. To address this need, we developed Librator, a system for analyzing and designing custom protein sequences of influenza virus hemagglutinin (HA) and neuraminidase (NA) glycoproteins. Within Librator's graphical interface, users can easily interrogate viral sequences and phylogenies, visualize antigen structures and conservation, mutate target residues and design custom antigens. Librator also provides optimized fragment design for Gibson Assembly of HA and NA expression constructs based on peptide conservation of all historical HA and NA sequences, ensuring fragments are reusable and compatible across related subtypes, thereby promoting reagent savings. Finally, the program facilitates single-cell immune profiling, epitope mapping of monoclonal antibodies and mosaic protein design. Using Librator-based antigen construction, we demonstrate that antigenicity can be readily transferred between HA molecules of H3, but not H1, lineage viruses. Altogether, Librator is a valuable tool for analyzing influenza virus HA and NA proteins and provides an efficient resource for optimizing recombinant influenza antigen synthesis.

With the rapid growth of high-resolution microscopy imaging data, revealing the subcellular map of human proteins has become a central task in the spatial proteome. The cell atlas of the Human Protein Atlas (HPA) provides precious resources for recognizing subcellular localization patterns at the cell level, and the large-scale annotated data enable learning via advanced deep neural networks. However, the existing predictors still suffer from the imbalanced class distribution and the lack of labeled data for minor classes. Thus, it is necessary to develop new methods for coping with these issues. We leverage the self-supervised learning protocol to address these problems. Especially, we propose a pre-training scheme to enhance the conventional supervised learning framework called SIFLoc. The pre-training is featured by a hybrid data augmentation method and a modified contrastive loss function, aiming to learn good feature representations from microscopic images. The experiments are performed on a large-scale immunofluorescence microscopic image dataset collected from the HPA database. Using the same deep neural networks as the classifier, the model pre-trained via SIFLoc not only outperforms the model without pre-training by a large margin but also shows advantages over the state-of-the-art self-supervised learning methods. Especially, SIFLoc improves the prediction accuracy for minor organelles significantly.

Detecting small and linked quantitative trait loci (QTLs) and QTL-by-environment interactions (QEIs) for complex traits is a difficult issue in immortalized F2 and F2:3 design, especially in the era of global climate change and environmental plasticity research. Here we proposed a compressed variance component mixed model. In this model, a parametric vector of QTL genotype and environment combination effects replaced QTL effects, environmental effects and their interaction effects, whereas the combination effect polygenic background replaced the QTL and QEI polygenic backgrounds. Thus, the number of variance components in the mixed model was greatly reduced. The model was incorporated into our genome-wide composite interval mapping (GCIM) to propose GCIM-QEI-random and GCIM-QEI-fixed, respectively, under random and fixed models of genetic effects. First, potentially associated QTLs and QEIs were selected from genome-wide scanning. Then, significant QTLs and QEIs were identified using empirical Bayes and likelihood ratio test. Finally, known and candidate genes around these significant loci were mined. The new methods were validated by a series of simulation studies and real data analyses. Compared with ICIM, GCIM-QEI-random had 29.77 ± 18.20% and 24.33 ± 10.15% higher average power, respectively, in 0.5-3.0% QTL and QEI detection, 43.44 ± 9.53% and 51.47 ± 15.70% higher average power, respectively, in linked QTL and QEI detection, and identified 30 more known genes for four rice yield traits, because GCIM-QEI-random identified more small genes/loci, being 2.69 ± 2.37% for additional genes. GCIM-QEI-random was slightly better than GCIM-QEI-fixed. In addition, the new methods may be extended into backcross and genome-wide association studies. This study provides effective methods for detecting small-effect and linked QTLs and QEIs.

Rapid development of single-cell RNA sequencing (scRNA-seq) technology has allowed researchers to explore biological phenomena at the cellular scale. Clustering is a crucial and helpful step for researchers to study the heterogeneity of cell. Although many clustering methods have been proposed, massive dropout events and the curse of dimensionality in scRNA-seq data make it still difficult to analysis because they reduce the accuracy of clustering methods, leading to misidentification of cell types. In this work, we propose the scHFC, which is a hybrid fuzzy clustering method optimized by natural computation based on Fuzzy C Mean (FCM) and Gath-Geva (GG) algorithms. Specifically, principal component analysis algorithm is utilized to reduce the dimensions of scRNA-seq data after it is preprocessed. Then, FCM algorithm optimized by simulated annealing algorithm and genetic algorithm is applied to cluster the data to output a membership matrix, which represents the initial clustering result and is taken as the input for GG algorithm to get the final clustering results. We also develop a cluster number estimation method called multi-index comprehensive estimation, which can estimate the cluster numbers well by combining four clustering effectiveness indexes. The performance of the scHFC method is evaluated on 17 scRNA-seq datasets, and compared with six state-of-the-art methods. Experimental results validate the better performance of our scHFC method in terms of clustering accuracy and stability of algorithm. In short, scHFC is an effective method to cluster cells for scRNA-seq data, and it presents great potential for downstream analysis of scRNA-seq data. The source code is available at https://github.com/WJ319/scHFC.

It's challenging work to identify disease-causing genes from the next-generation sequencing (NGS) data of patients with Mendelian disorders. To improve this situation, researchers have developed many phenotype-driven gene prioritization methods using a patient's genotype and phenotype information, or phenotype information only as input to rank the candidate's pathogenic genes. Evaluations of these ranking methods provide practitioners with convenience for choosing an appropriate tool for their workflows, but retrospective benchmarks are underpowered to provide statistically significant results in their attempt to differentiate. In this research, the performance of ten recognized causal-gene prioritization methods was benchmarked using 305 cases from the Deciphering Developmental Disorders (DDD) project and 209 in-house cases via a relatively unbiased methodology. The evaluation results show that methods using Human Phenotype Ontology (HPO) terms and Variant Call Format (VCF) files as input achieved better overall performance than those using phenotypic data alone. Besides, LIRICAL and AMELIE, two of the best methods in our benchmark experiments, complement each other in cases with the causal genes ranked highly, suggesting a possible integrative approach to further enhance the diagnostic efficiency. Our benchmarking provides valuable reference information to the computer-assisted rapid diagnosis in Mendelian diseases and sheds some light on the potential direction of future improvement on disease-causing gene prioritization methods.

Targeted drugs have been applied to the treatment of cancer on a large scale, and some patients have certain therapeutic effects. It is a time-consuming task to detect drug-target interactions (DTIs) through biochemical experiments. At present, machine learning (ML) has been widely applied in large-scale drug screening. However, there are few methods for multiple information fusion. We propose a multiple kernel-based triple collaborative matrix factorization (MK-TCMF) method to predict DTIs. The multiple kernel matrices (contain chemical, biological and clinical information) are integrated via multi-kernel learning (MKL) algorithm. And the original adjacency matrix of DTIs could be decomposed into three matrices, including the latent feature matrix of the drug space, latent feature matrix of the target space and the bi-projection matrix (used to join the two feature spaces). To obtain better prediction performance, MKL algorithm can regulate the weight of each kernel matrix according to the prediction error. The weights of drug side-effects and target sequence are the highest. Compared with other computational methods, our model has better performance on four test data sets.

Erratum in Brief Bioinform. 2022 Mar 02;:

Sequence logos are used to visually display conservations and variations in short sequences. They can indicate the fixed patterns or conserved motifs in a batch of DNA or protein sequences. However, most of the popular sequence logo generators are based on the assumption that all the input sequences are from the same homologous group, which will lead to an overlook of the heterogeneity among the sequences during the sequence logo making process. Heterogeneous groups of sequences may represent clades of different evolutionary origins, or genes families with different functions. Therefore, it is essential to divide the sequences into different phylogenetic or functional groups to reveal their specific sequence motifs and conservation patterns. To solve these problems, we developed MetaLogo, which can automatically cluster the input sequences after multiple sequence alignment and phylogenetic tree construction, and then output sequence logos for multiple groups and aligned them in one figure. User-defined grouping is also supported by MetaLogo to allow users to investigate functional motifs in a more delicate and dynamic perspective. MetaLogo can highlight both the homologous and nonhomologous sites among sequences. MetaLogo can also be used to annotate the evolutionary positions and gene functions of unknown sequences, together with their local sequence characteristics. We provide users a public MetaLogo web server (http://metalogo.omicsnet.org), a standalone Python package (https://github.com/labomics/MetaLogo), and also a built-in web server available for local deployment. Using MetaLogo, users can draw informative, customized and publishable sequence logos without any programming experience to present and investigate new knowledge on specific sequence sets.

Bacterial genomes are massively sequenced, and they provide valuable data to better know the complete set of genes of a species. The analysis of thousands of bacterial strains can identify both shared genes and those appearing only in the pathogenic ones. Current computational gene finders facilitate this task but often miss some existing genes. However, the present availability of different genomes from the same species is useful to estimate the selective pressure applied on genes of complete pangenomes. It may assist in evaluating gene predictions either by checking the certainty of a new gene or annotating it as a gene under positive selection. Here, we estimated the selective pressure of 19 271 genes that are part of the pangenome of the human opportunistic pathogen Acinetobacter baumannii and found that most genes in this bacterium are subject to negative selection. However, 23% of them showed values compatible with positive selection. These latter were mainly uncharacterized proteins or genes required to evade the host defence system including genes related to resistance and virulence whose changes may be favoured to acquire new functions. Finally, we evaluated the utility of measuring selection pressure in the detection of sequencing errors and the validation of gene prediction.

DNA sequencing technologies have advanced significantly in the last few years leading to advancements in biomedical research which has improved personalised medicine and the discovery of new treatments for diseases. Sequencing technology advancement has also reduced the cost of DNA sequencing, which has led to the rise of direct-to-consumer (DTC) sequencing, e.g. 23andme.com, ancestry.co.uk, etc. In the meantime, concerns have emerged over privacy and security in collecting, handling, analysing and sharing DNA and genomic data. DNA data are unique and can be used to identify individuals. Moreover, those data provide information on people's current disease status and disposition, e.g. mental health or susceptibility for developing cancer. DNA privacy violation does not only affect the owner but also affects their close consanguinity due to its hereditary nature. This article introduces and defines the term 'digital DNA life cycle' and presents an overview of privacy and security threats and their mitigation techniques for predigital DNA and throughout the digital DNA life cycle. It covers DNA sequencing hardware, software and DNA sequence pipeline in addition to common privacy attacks and their countermeasures when DNA digital data are stored, queried or shared. Likewise, the article examines DTC genomic sequencing privacy and security.

MOTIVATION: Protein-protein interactions (PPIs) are central to most biological processes. However, reliable identification of PPI sites using conventional experimental methods is slow and expensive. Therefore, great efforts are being put into computational methods to identify PPI sites. RESULTS: We present Edge Aggregated GRaph Attention NETwork (EGRET), a highly accurate deep learning-based method for PPI site prediction, where we have used an edge aggregated graph attention network to effectively leverage the structural information. We, for the first time, have used transfer learning in PPI site prediction. Our proposed edge aggregated network, together with transfer learning, has achieved notable improvement over the best alternate methods. Furthermore, we systematically investigated EGRET's network behavior to provide insights about the causes of its decisions. AVAILABILITY: EGRET is freely available as an open source project at https://github.com/Sazan-Mahbub/EGRET. CONTACT: shams_bayzid@cse.buet.ac.bd.

With the rapid accumulation of microbiome data around the world, numerous computational bioinformatics methods have been developed for pattern mining from such paramount microbiome data. Current microbiome data mining methods, such as gene and species mining, rely heavily on sequence comparison. Most of these methods, however, have a clear trade-off, particularly, when it comes to big-data analytical efficiency and accuracy. Microbiome entities are usually organized in ontology structures, and pattern mining methods that have considered ontology structures could offer advantages in mining efficiency and accuracy. Here, we have summarized the ontology-aware neural network (ONN) as a novel framework for microbiome data mining. We have discussed the applications of ONN in multiple contexts, including gene mining, species mining and microbial community dynamic pattern mining. We have then highlighted one of the most important characteristics of ONN, namely, novel knowledge discovery, which makes ONN a standout among all microbiome data mining methods. Finally, we have provided several applications to showcase the advantage of ONN over other methods in microbiome data mining. In summary, ONN represents a paradigm shift for pattern mining from microbiome data: from traditional machine learning approach to ontology-aware and model-based approach, which has found its broad application scenarios in microbiome data mining.

Biomedical data are becoming increasingly multimodal and thereby capture the underlying complex relationships among biological processes. Deep learning (DL)-based data fusion strategies are a popular approach for modeling these nonlinear relationships. Therefore, we review the current state-of-the-art of such methods and propose a detailed taxonomy that facilitates more informed choices of fusion strategies for biomedical applications, as well as research on novel methods. By doing so, we find that deep fusion strategies often outperform unimodal and shallow approaches. Additionally, the proposed subcategories of fusion strategies show different advantages and drawbacks. The review of current methods has shown that, especially for intermediate fusion strategies, joint representation learning is the preferred approach as it effectively models the complex interactions of different levels of biological organization. Finally, we note that gradual fusion, based on prior biological knowledge or on search strategies, is a promising future research path. Similarly, utilizing transfer learning might overcome sample size limitations of multimodal data sets. As these data sets become increasingly available, multimodal DL approaches present the opportunity to train holistic models that can learn the complex regulatory dynamics behind health and disease.

Open chromatin regions (OCRs) allow direct interaction between cis-regulatory elements and trans-acting factors. Therefore, predicting all potential OCR-mediated loops is essential for deciphering the regulation mechanism of gene expression. However, existing loop prediction tools are restricted to specific anchor types. Here, we present CharID (Chromatin Accessible Region Interaction Detector), a two-step model that combines neural network and ensemble learning to predict OCR-mediated loops. In the first step, CharID-Anchor, an attention-based hybrid CNN-BiGRU network is constructed to discriminate between the anchor and nonanchor OCRs. In the second step, CharID-Loop uses gradient boosting decision tree with chromosome-split strategy to predict the interactions between anchor OCRs. The performance was assessed in three human cell lines, and CharID showed superior prediction performance compared with other algorithms. In contrast to the methods designed to predict a particular type of loops, CharID can detect varieties of chromatin loops not limited to enhancer-promoter loops or architectural protein-mediated loops. We constructed the OCR-mediated interaction network using the predicted loops and identified hub anchors, which are highlighted by their proximity to housekeeping genes. By analyzing loops containing SNPs associated with cardiovascular disease, we identified an SNP-gene loop indicating the regulation mechanism of the GFOD1. Taken together, CharID universally predicts diverse chromatin loops beyond other state-of-the-art methods, which are limited by anchor types, and experimental techniques, which are limited by sensitivities drastically decaying with the genomic distance of anchors. Finally, we hosted Peaksniffer, a user-friendly web server that provides online prediction, query and visualization of OCRs and associated loops.

A transcriptome constructed from short-read RNA sequencing (RNA-seq) is an easily attainable proxy catalog of protein-coding genes when genome assembly is unnecessary, expensive or difficult. In the absence of a sequenced genome to guide the reconstruction process, the transcriptome must be assembled de novo using only the information available in the RNA-seq reads. Subsequently, the sequences must be annotated in order to identify sequence-intrinsic and evolutionary features in them (for example, protein-coding regions). Although straightforward at first glance, de novo transcriptome assembly and annotation can quickly prove to be challenging undertakings. In addition to familiarizing themselves with the conceptual and technical intricacies of the tasks at hand and the numerous pre- and post-processing steps involved, those interested must also grapple with an overwhelmingly large choice of tools. The lack of standardized workflows, fast pace of development of new tools and techniques and paucity of authoritative literature have served to exacerbate the difficulty of the task even further. Here, we present a comprehensive overview of de novo transcriptome assembly and annotation. We discuss the procedures involved, including pre- and post-processing steps, and present a compendium of corresponding tools.

In the past few decades, chronic hepatitis B caused by hepatitis B virus (HBV) has been one of the most serious diseases to human health. The development of innovative systems is essential for preventing the complex pathogenesis of hepatitis B and reducing side effects caused by drugs. HBV inhibitory drugs have been developed through various compounds, and they are often limited by routine experimental screening and delay drug development. More recently, virtual screening of compounds has gradually been used in drug research with strong computational capability and is further applied in anti-HBV drug screening, thus facilitating a reliable drug screening process. However, the lack of structural information in traditional compound analysis is an important hurdle for unsatisfactory efficiency in drug screening. Here, a natural language processing technique was adopted to analyze compound simplified molecular input line entry system strings. By using the targeted optimized word2vec model for pretraining, we can accurately represent the relationship between the compound and its substructure. The machine learning model based on training results can effectively predict the inhibitory effect of compounds on HBV and liver toxicity. The reliability of the model is verified by the results of wet-lab experiments. In addition, a tool has been published to predict potential compounds. Hence, this article provides a new perspective on the prediction of compound properties for anti-HBV drugs that can help improve hepatitis B diagnosis and further develop human health in the future.

A critical issue of drug risk-benefit evaluation is to determine the frequencies of drug side effects. Randomized controlled trail is the conventional method for obtaining the frequencies of side effects, while it is laborious and slow. Therefore, it is necessary to guide the trail by computational methods. Existing methods for predicting the frequencies of drug side effects focus on modeling drug-side effect interaction graph. The inherent disadvantage of these approaches is that their performance is closely linked to the density of interactions but which is highly sparse. More importantly, for a cold start drug that does not appear in the training data, such methods cannot learn the preference embedding of the drug because there is no link to the drug in the interaction graph. In this work, we propose a new method for predicting the frequencies of drug side effects, DSGAT, by using the drug molecular graph instead of the commonly used interaction graph. This leads to the ability to learn embeddings for cold start drugs with graph attention networks. The proposed novel loss function, i.e. weighted $\varepsilon$-insensitive loss function, could alleviate the sparsity problem. Experimental results on one benchmark dataset demonstrate that DSGAT yields significant improvement for cold start drugs and outperforms the state-of-the-art performance in the warm start scenario. Source code and datasets are available at https://github.com/xxy45/DSGAT.

Correctly identifying the true driver mutations in a patient's tumor is a major challenge in precision oncology. Most efforts address frequent mutations, leaving medium- and low-frequency variants mostly unaddressed. For TP53, this identification is crucial for both somatic and germline mutations, with the latter associated with the Li-Fraumeni syndrome (LFS), a multiorgan cancer predisposition. We present TP53_PROF (prediction of functionality), a gene specific machine learning model to predict the functional consequences of every possible missense mutation in TP53, integrating human cell- and yeast-based functional assays scores along with computational scores. Variants were labeled for the training set using well-defined criteria of prevalence in four cancer genomics databases. The model's predictions provided accuracy of 96.5%. They were validated experimentally, and were compared to population data, LFS datasets, ClinVar annotations and to TCGA survival data. Very high accuracy was shown through all methods of validation. TP53_PROF allows accurate classification of TP53 missense mutations applicable for clinical practice. Our gene specific approach integrated machine learning, highly reliable features and biological knowledge, to create an unprecedented, thoroughly validated and clinically oriented classification model. This approach currently addresses TP53 mutations and will be applied in the future to other important cancer genes.

Predicting the response of cancer patients to a particular treatment is a major goal of modern oncology and an important step toward personalized treatment. In the practical clinics, the clinicians prefer to obtain the most-suited drugs for a particular patient instead of knowing the exact values of drug sensitivity. Instead of predicting the exact value of drug response, we proposed a deep learning-based method, named Siamese Response Deep Factorization Machines (SRDFM) Network, for personalized anti-cancer drug recommendation, which directly ranks the drugs and provides the most effective drugs. A Siamese network (SN), a type of deep learning network that is composed of identical subnetworks that share the same architecture, parameters and weights, was used to measure the relative position (RP) between drugs for each cell line. Through minimizing the difference between the real RP and the predicted RP, an optimal SN model was established to provide the rank for all the candidate drugs. Specifically, the subnetwork in each side of the SN consists of a feature generation level and a predictor construction level. On the feature generation level, both drug property and gene expression, were adopted to build a concatenated feature vector, which even enables the recommendation for newly designed drugs with only chemical property known. Particularly, we developed a response unit here to generate weighted genetic feature vector to simulate the biological interaction mechanism between a specific drug and the genes. For the predictor construction level, we built this level integrating a factorization machine (FM) component with a deep neural network component. The FM can well handle the discrete chemical information and both low-order and high-order feature interactions could be sufficiently learned. Impressively, the SRDFM works well on both single-drug recommendation and synergic drug combination. Experiment result on both single-drug and synergetic drug data sets have shown the efficiency of the SRDFM. The Python implementation for the proposed SRDFM is available at at https://github.com/RanSuLab/SRDFM Contact: ran.su@tju.edu.cn, gbx@mju.edu.cn and weileyi@sdu.edu.cn.

Identification of adverse drug events (ADEs) is crucial to reduce human health risks and improve drug safety assessment. With an increasing number of biological and medical data, computational methods such as network-based methods were proposed for ADE prediction with high efficiency and low cost. However, previous network-based methods rely on the topological information of known drug-ADE networks, and hence cannot make predictions for novel compounds without any known ADE. In this study, we introduced chemical substructures to bridge the gap between the drug-ADE network and novel compounds, and developed a novel network-based method named ADENet, which can predict potential ADEs for not only drugs within the drug-ADE network, but also novel compounds outside the network. To show the performance of ADENet, we collected drug-ADE associations from a comprehensive database named MetaADEDB and constructed a series of network-based prediction models. These models obtained high area under the receiver operating characteristic curve values ranging from 0.871 to 0.947 in 10-fold cross-validation. The best model further showed high performance in external validation, which outperformed a previous network-based and a recent deep learning-based method. Using several approved drugs as case studies, we found that 32-54% of the predicted ADEs can be validated by the literature, indicating the practical value of ADENet. Moreover, ADENet is freely available at our web server named NetInfer (http://lmmd.ecust.edu.cn/netinfer). In summary, our method would provide a promising tool for ADE prediction and drug safety assessment in drug discovery and development.

Annotation of cells in single-cell clustering requires a homogeneous grouping of cell populations. Since single-cell data are susceptible to technical noise, the quality of genes selected prior to clustering is of crucial importance in the preliminary steps of downstream analysis. Therefore, interest in robust gene selection has gained considerable attention in recent years. We introduce sc-REnF [robust entropy based feature (gene) selection method], aiming to leverage the advantages of $R{\prime}{e}nyi$ and $Tsallis$ entropies in gene selection for single cell clustering. Experiments demonstrate that with tuned parameter ($q$), $R{\prime}{e}nyi$ and $Tsallis$ entropies select genes that improved the clustering results significantly, over the other competing methods. sc-REnF can capture relevancy and redundancy among the features of noisy data extremely well due to its robust objective function. Moreover, the selected features/genes can able to determine the unknown cells with a high accuracy. Finally, sc-REnF yields good clustering performance in small sample, large feature scRNA-seq data. Availability: The sc-REnF is available at https://github.com/Snehalikalall/sc-REnF.

Direct coupling analysis (DCA) has been widely used to infer evolutionary coupled residue pairs from the multiple sequence alignment (MSA) of homologous sequences. However, effectively selecting residue pairs with significant evolutionary couplings according to the result of DCA is a non-trivial task. In this study, we developed a general statistical framework for significant evolutionary coupling detection, referred to as irreproducible discovery rate (IDR)-DCA, which is based on reproducibility analysis of the coupling scores obtained from DCA on manually created MSA replicates. IDR-DCA was applied to select residue pairs for contact prediction for monomeric proteins, protein-protein interactions and monomeric RNAs, in which three different versions of DCA were applied. We demonstrated that with the application of IDR-DCA, the residue pairs selected using a universal threshold always yielded stable performance for contact prediction. Comparing with the application of carefully tuned coupling score cutoffs, IDR-DCA always showed better performance. The robustness of IDR-DCA was also supported through the MSA downsampling analysis. We further demonstrated the effectiveness of applying constraints obtained from residue pairs selected by IDR-DCA to assist RNA secondary structure prediction.

Optimal methods could effectively improve the accuracy of predicting and identifying candidate driver genes. Various computational methods based on mutational frequency, network and function approaches have been developed to identify mutation driver genes in cancer genomes. However, a comprehensive evaluation of the performance levels of network-, function- and frequency-based methods is lacking. In the present study, we assessed and compared eight performance criteria for eight network-based, one function-based and three frequency-based algorithms using eight benchmark datasets. Under different conditions, the performance of approaches varied in terms of network, measurement and sample size. The frequency-based driverMAPS and network-based HotNet2 methods showed the best overall performance. Network-based algorithms using protein-protein interaction networks outperformed the function- and the frequency-based approaches. Precision, F1 score and Matthews correlation coefficient were low for most approaches. Thus, most of these algorithms require stringent cutoffs to correctly distinguish driver and non-driver genes. We constructed a website named Cancer Driver Catalog (http://159.226.67.237/sun/cancer_driver/), wherein we integrated the gene scores predicted by the foregoing software programs. This resource provides valuable guidance for cancer researchers and clinical oncologists prioritizing cancer driver gene candidates by using an optimal tool.

The cell type identification is among the most important tasks in single-cell RNA-sequencing (scRNA-seq) analysis. Many in silico methods have been developed and can be roughly categorized as either supervised or unsupervised. In this study, we investigated the performances of 8 supervised and 10 unsupervised cell type identification methods using 14 public scRNA-seq datasets of different tissues, sequencing protocols and species. We investigated the impacts of a number of factors, including total amount of cells, number of cell types, sequencing depth, batch effects, reference bias, cell population imbalance, unknown/novel cell type, and computational efficiency and scalability. Instead of merely comparing individual methods, we focused on factors' impacts on the general category of supervised and unsupervised methods. We found that in most scenarios, the supervised methods outperformed the unsupervised methods, except for the identification of unknown cell types. This is particularly true when the supervised methods use a reference dataset with high informational sufficiency, low complexity and high similarity to the query dataset. However, such outperformance could be undermined by some undesired dataset properties investigated in this study, which lead to uninformative and biased reference datasets. In these scenarios, unsupervised methods could be comparable to supervised methods. Our study not only explained the cell typing methods' behaviors under different experimental settings but also provided a general guideline for the choice of method according to the scientific goal and dataset properties. Finally, our evaluation workflow is implemented as a modularized R pipeline that allows future evaluation of new methods. Availability: All the source codes are available at https://github.com/xsun28/scRNAIdent.

Location proteomics seeks to provide automated high-resolution descriptions of protein location patterns within cells. Many efforts have been undertaken in location proteomics over the past decades, thereby producing plenty of automated predictors for protein subcellular localization. However, most of these predictors are trained solely from high-throughput microscopic images or protein amino acid sequences alone. Unifying heterogeneous protein data sources has yet to be exploited. In this paper, we present a pipeline called sequence, image, network-based protein subcellular locator (SIN-Locator) that constructs a multi-view description of proteins by integrating multiple data types including images of protein expression in cells or tissues, amino acid sequences and protein-protein interaction networks, to classify the patterns of protein subcellular locations. Proteins were encoded by both handcrafted features and deep learning features, and multiple combining methods were implemented. Our experimental results indicated that optimal integrations can considerately enhance the classification accuracy, and the utility of SIN-Locator has been demonstrated through applying to new released proteins in the human protein atlas. Furthermore, we also investigate the contribution of different data sources and influence of partial absence of data. This work is anticipated to provide clues for reconciliation and combination of multi-source data for protein location analysis.

Large metabolomics datasets inevitably contain unwanted technical variations which can obscure meaningful biological signals and affect how this information is applied to personalized healthcare. Many methods have been developed to handle unwanted variations. However, the underlying assumptions of many existing methods only hold for a few specific scenarios. Some tools remove technical variations with models trained on quality control (QC) samples which may not generalize well on subject samples. Additionally, almost none of the existing methods supports datasets with multiple types of QC samples, which greatly limits their performance and flexibility. To address these issues, a non-parametric method TIGER (Technical variation elImination with ensemble learninG architEctuRe) is developed in this study and released as an R package (https://CRAN.R-project.org/package=TIGERr). TIGER integrates the random forest algorithm into an adaptable ensemble learning architecture. Evaluation results show that TIGER outperforms four popular methods with respect to robustness and reliability on three human cohort datasets constructed with targeted or untargeted metabolomics data. Additionally, a case study aiming to identify age-associated metabolites is performed to illustrate how TIGER can be used for cross-kit adjustment in a longitudinal analysis with experimental data of three time-points generated by different analytical kits. A dynamic website is developed to help evaluate the performance of TIGER and examine the patterns revealed in our longitudinal analysis (https://han-siyu.github.io/TIGER_web/). Overall, TIGER is expected to be a powerful tool for metabolomics data analysis.

Accurate cell classification is the groundwork for downstream analysis of single-cell sequencing data, yet how to identify true marker genes for different cell types still remains a big challenge. Here, we report COSine similarity-based marker Gene identification (COSG) as a cosine similarity-based method for more accurate and scalable marker gene identification. COSG is applicable to single-cell RNA sequencing data, single-cell ATAC sequencing data and spatially resolved transcriptome data. COSG is fast and scalable for ultra-large datasets of million-scale cells. Application on both simulated and real experimental datasets showed that the marker genes or genomic regions identified by COSG have greater cell-type specificity, demonstrating the superior performance of COSG in terms of both accuracy and efficiency as compared with other available methods.

Messenger RNA (mRNA) vaccines have shown great potential for anti-tumor therapy due to the advantages in safety, efficacy and industrial production. However, it remains a challenge to identify suitable cancer neoantigens that can be targeted for mRNA vaccines. Abnormal alternative splicing occurs in a variety of tumors, which may result in the translation of abnormal transcripts into tumor-specific proteins. High-throughput technologies make it possible for systematic characterization of alternative splicing as a source of suitable target neoantigens for mRNA vaccine development. Here, we summarized difficulties and challenges for identifying alternative splicing-derived cancer neoantigens from RNA-seq data and proposed a conceptual framework for designing personalized mRNA vaccines based on alternative splicing-derived cancer neoantigens. In addition, several points were presented to spark further discussion toward improving the identification of alternative splicing-derived cancer neoantigens.

Single nucleotide polymorphisms (SNPs) are the most abundant type of genomic variation and the most accessible to genotype in large cohorts. However, they individually explain a small proportion of phenotypic differences between individuals. Ancestry, collective SNP effects, structural variants, somatic mutations or even differences in historic recombination can potentially explain a high percentage of genomic divergence. These genetic differences can be infrequent or laborious to characterize; however, many of them leave distinctive marks on the SNPs across the genome allowing their study in large population samples. Consequently, several methods have been developed over the last decade to detect and analyze different genomic structures using SNP arrays, to complement genome-wide association studies and determine the contribution of these structures to explain the phenotypic differences between individuals. We present an up-to-date collection of available bioinformatics tools that can be used to extract relevant genomic information from SNP array data including population structure and ancestry; polygenic risk scores; identity-by-descent fragments; linkage disequilibrium; heritability and structural variants such as inversions, copy number variants, genetic mosaicisms and recombination histories. From a systematic review of recently published applications of the methods, we describe the main characteristics of R packages, command-line tools and desktop applications, both free and commercial, to help make the most of a large amount of publicly available SNP data.

Polygenic scores (PGS) are important tools for carrying out genetic prediction of common diseases and disease related complex traits, facilitating the development of precision medicine. Unfortunately, despite the critical importance of PGS and the vast number of PGS methods recently developed, few comprehensive comparison studies have been performed to evaluate the effectiveness of PGS methods. To fill this critical knowledge gap, we performed a comprehensive comparison study on 12 different PGS methods through internal evaluations on 25 quantitative and 25 binary traits within the UK Biobank with sample sizes ranging from 147 408 to 336 573, and through external evaluations via 25 cross-study and 112 cross-ancestry analyses on summary statistics from multiple genome-wide association studies with sample sizes ranging from 1415 to 329 345. We evaluate the prediction accuracy, computational scalability, as well as robustness and transferability of different PGS methods across datasets and/or genetic ancestries, providing important guidelines for practitioners in choosing PGS methods. Besides method comparison, we present a simple aggregation strategy that combines multiple PGS from different methods to take advantage of their distinct benefits to achieve stable and superior prediction performance. To facilitate future applications of PGS, we also develop a PGS webserver (http://www.pgs-server.com/) that allows users to upload summary statistics and choose different PGS methods to fit the data directly. We hope that our results, method and webserver will facilitate the routine application of PGS across different research areas.

Identifying genome-wide binding events between circular RNAs (circRNAs) and RNA-binding proteins (RBPs) can greatly facilitate our understanding of functional mechanisms within circRNAs. Thanks to the development of cross-linked immunoprecipitation sequencing technology, large amounts of genome-wide circRNA binding event data have accumulated, providing opportunities for designing high-performance computational models to discriminate RBP interaction sites and thus to interpret the biological significance of circRNAs. Unfortunately, there are still no computational models sufficiently flexible to accommodate circRNAs from different data scales and with various degrees of feature representation. Here, we present HCRNet, a novel end-to-end framework for identification of circRNA-RBP binding events. To capture the hierarchical relationships, the multi-source biological information is fused to represent circRNAs, including various natural language sequence features. Furthermore, a deep temporal convolutional network incorporating global expectation pooling was developed to exploit the latent nucleotide dependencies in an exhaustive manner. We benchmarked HCRNet on 37 circRNA datasets and 31 linear RNA datasets to demonstrate the effectiveness of our proposed method. To evaluate further the model's robustness, we performed HCRNet on a full-length dataset containing 740 circRNAs. Results indicate that HCRNet generally outperforms existing methods. In addition, motif analyses were conducted to exhibit the interpretability of HCRNet on circRNAs. All supporting source code and data can be downloaded from https://github.com/yangyn533/HCRNet and https://doi.org/10.6084/m9.figshare.16943722.v1. And the web server of HCRNet is publicly accessible at http://39.104.118.143:5001/.

Changes in protein sequence can have dramatic effects on how proteins fold, their stability and dynamics. Over the last 20 years, pioneering methods have been developed to try to estimate the effects of missense mutations on protein stability, leveraging growing availability of protein 3D structures. These, however, have been developed and validated using experimentally derived structures and biophysical measurements. A large proportion of protein structures remain to be experimentally elucidated and, while many studies have based their conclusions on predictions made using homology models, there has been no systematic evaluation of the reliability of these tools in the absence of experimental structural data. We have, therefore, systematically investigated the performance and robustness of ten widely used structural methods when presented with homology models built using templates at a range of sequence identity levels (from 15% to 95%) and contrasted performance with sequence-based tools, as a baseline. We found there is indeed performance deterioration on homology models built using templates with sequence identity below 40%, where sequence-based tools might become preferable. This was most marked for mutations in solvent exposed residues and stabilizing mutations. As structure prediction tools improve, the reliability of these predictors is expected to follow, however we strongly suggest that these factors should be taken into consideration when interpreting results from structure-based predictors of mutation effects on protein stability.

In recent years, increasing biological experiments and scientific studies have demonstrated that microRNA (miRNA) plays an important role in the development of human complex diseases. Therefore, discovering miRNA-disease associations can contribute to accurate diagnosis and effective treatment of diseases. Identifying miRNA-disease associations through computational methods based on biological data has been proven to be low-cost and high-efficiency. In this study, we proposed a computational model named Stacked Autoencoder for potential MiRNA-Disease Association prediction (SAEMDA). In SAEMDA, all the miRNA-disease samples were used to pretrain a Stacked Autoencoder (SAE) in an unsupervised manner. Then, the positive samples and the same number of selected negative samples were utilized to fine-tune SAE in a supervised manner after adding an output layer with softmax classifier to the SAE. SAEMDA can make full use of the feature information of all unlabeled miRNA-disease pairs. Therefore, SAEMDA is suitable for our dataset containing small labeled samples and large unlabeled samples. As a result, SAEMDA achieved AUCs of 0.9210 and 0.8343 in global and local leave-one-out cross validation. Besides, SAEMDA obtained an average AUC and standard deviation of 0.9102 ± /-0.0029 in 100 times of 5-fold cross validation. These results were better than those of previous models. Moreover, we carried out three case studies to further demonstrate the predictive accuracy of SAEMDA. As a result, 82% (breast neoplasms), 100% (lung neoplasms) and 90% (esophageal neoplasms) of the top 50 predicted miRNAs were verified by databases. Thus, SAEMDA could be a useful and reliable model to predict potential miRNA-disease associations.

Phosphorylation of proteins is one of the most significant post-translational modifications (PTMs) and plays a crucial role in plant functionality due to its impact on signaling, gene expression, enzyme kinetics, protein stability and interactions. Accurate prediction of plant phosphorylation sites (p-sites) is vital as abnormal regulation of phosphorylation usually leads to plant diseases. However, current experimental methods for PTM prediction suffers from high-computational cost and are error-prone. The present study develops machine learning-based prediction techniques, including a high-performance interpretable deep tabular learning network (TabNet) to improve the prediction of protein p-sites in soybean. Moreover, we use a hybrid feature set of sequential-based features, physicochemical properties and position-specific scoring matrices to predict serine (Ser/S), threonine (Thr/T) and tyrosine (Tyr/Y) p-sites in soybean for the first time. The experimentally verified p-sites data of soybean proteins are collected from the eukaryotic phosphorylation sites database and database post-translational modification. We then remove the redundant set of positive and negative samples by dropping protein sequences with >40% similarity. It is found that the developed techniques perform >70% in terms of accuracy. The results demonstrate that the TabNet model is the best performing classifier using hybrid features and with window size of 13, resulted in 78.96 and 77.24% sensitivity and specificity, respectively. The results indicate that the TabNet method has advantages in terms of high-performance and interpretability. The proposed technique can automatically analyze the data without any measurement errors and any human intervention. Furthermore, it can be used to predict putative protein p-sites in plants effectively. The collected dataset and source code are publicly deposited at https://github.com/Elham-khalili/Soybean-P-sites-Prediction.

With the rapid progress of deep learning in cryo-electron microscopy and protein structure prediction, improving the accuracy of the protein structure model by using a density map and predicted contact/distance map through deep learning has become an urgent need for robust methods. Thus, designing an effective protein structure optimization strategy based on the density map and predicted contact/distance map is critical to improving the accuracy of structure refinement. In this article, a protein structure optimization method based on the density map and predicted contact/distance map by deep-learning technology was proposed in accordance with the result of matching between the density map and the initial model. Physics- and knowledge-based energy functions, integrated with Cryo-EM density map data and deep-learning data, were used to optimize the protein structure in the simulation. The dynamic confidence score was introduced to the iterative process for choosing whether it is a density map or a contact/distance map to dominate the movement in the simulation to improve the accuracy of refinement. The protocol was tested on a large set of 224 non-homologous membrane proteins and generated 214 structural models with correct folds, where 4.5% of structural models were generated from structural models with incorrect folds. Compared with other state-of-the-art methods, the major advantage of the proposed methods lies in the skills for using density map and contact/distance map in the simulation, as well as the new energy function in the re-assembly simulations. Overall, the results demonstrated that this strategy is a valuable approach and ready to use for atomic-level structure refinement using cryo-EM density map and predicted contact/distance map.

Shotgun sequencing is routinely employed to study bacteria in microbial communities. With the vast amount of shotgun sequencing reads generated in a metagenomic project, it is crucial to determine the microbial composition at the strain level. This study investigated 20 computational tools that attempt to infer bacterial strain genomes from shotgun reads. For the first time, we discussed the methodology behind these tools. We also systematically evaluated six novel-strain-targeting tools on the same datasets and found that BHap, mixtureS and StrainFinder performed better than other tools. Because the performance of the best tools is still suboptimal, we discussed future directions that may address the limitations.

In recent decades, exploring potential relationships between diseases has been an active research field. With the rapid accumulation of disease-related biomedical data, a lot of computational methods and tools/platforms have been developed to reveal intrinsic relationship between diseases, which can provide useful insights to the study of complex diseases, e.g. understanding molecular mechanisms of diseases and discovering new treatment of diseases. Human complex diseases involve both external phenotypic abnormalities and complex internal molecular mechanisms in organisms. Computational methods with different types of biomedical data from phenotype to genotype can evaluate disease-disease associations at different levels, providing a comprehensive perspective for understanding diseases. In this review, available biomedical data and databases for evaluating disease-disease associations are first summarized. Then, existing computational methods for disease-disease associations are reviewed and classified into five groups in terms of the usages of biomedical data, including disease semantic-based, phenotype-based, function-based, representation learning-based and text mining-based methods. Further, we summarize software tools/platforms for computation and analysis of disease-disease associations. Finally, we give a discussion and summary on the research of disease-disease associations. This review provides a systematic overview for current disease association research, which could promote the development and applications of computational methods and tools/platforms for disease-disease associations.

In this paper, we study the problem for finding complex proteoforms from protein databases based on top-down tandem mass spectrum data. The main difficulty to solve the problem is to handle the combinatorial explosion of various alterations on a protein. To overcome the combinatorial explosion of various alterations on a protein, the problem has been formulated as the alignment problem of a proteoform mass graph (PMG) and a spectrum mass graph (SMG). The other important issue is to handle mass errors of peaks in the input spectrum. In previous methods, an error tolerance value is used to handle the mass differences between the matched consecutive nodes/peaks in PMG and SMG. However, such a way to handle mass error can not guarantee that the mass difference between any pairs of nodes in the alignment is approximately the same for both PMG and SMG. It may lead to large error accumulation if positive (or negative) errors occur consecutively for a large number of consecutive matched node pairs. The problem is severe so that some existing software packages include a step to further refine the alignments. In this paper, we propose a new model to handle the mass errors of peaks based on the formulation of the PMG and SMG. Note that the masses of sub-paths on the PMG are theoretical and suppose to be accurate. Our method allows each peak in the input spectrum to have a predefined error range. In the alignment of PMG and SMG, we need to give a correction of the mass for each matched peak within the predefined error range. After the correction, we impose that the mass between any two (not necessarily consecutive) matched nodes in the PMG is identical to that of the corresponding two matched peaks in the SMG. Intuitively, this kind of alignment is more accurate. We design an algorithm to find a maximum number of matched node and peak pairs in the two (PMG and SMG) mass graphs under the new constraint. The obtained alignment can show matched node and peak pairs as well as the corrected positions of peaks. The algorithm works well for moderate size input instances and takes very long time as well as huge size memory for large input size instances. Therefore, we propose an algorithm to do diagonal alignment. The diagonal alignment algorithm can solve large input size instances in reasonable time. Experiments show that our new algorithms can report alignments with much larger number of matched node pairs. The software package and test data sets are available at https://github.com/Zeirdo/TopMGRefine.

Numerous cancer types have shown to present hypermethylation of CpG islands, also known as a CpG island methylator phenotype (CIMP), often associated with survival variation. Despite extensive research on CIMP, the etiology of this variability remains elusive, possibly due to lack of consistency in defining CIMP. In this work, we utilize a pan-cancer approach to further explore CIMP, focusing on 26 cancer types profiled in the Cancer Genome Atlas (TCGA). We defined CIMP systematically and agnostically, discarding any effects associated with age, gender or tumor purity. We then clustered samples based on their most variable DNA methylation values and analyzed resulting patient groups. Our results confirmed the existence of CIMP in 19 cancers, including gliomas and colorectal cancer. We further showed that CIMP was associated with survival differences in eight cancer types and, in five, represented a prognostic biomarker independent of clinical factors. By analyzing genetic and transcriptomic data, we further uncovered potential drivers of CIMP and classified them in four categories: mutations in genes directly involved in DNA demethylation; mutations in histone methyltransferases; mutations in genes not involved in methylation turnover, such as KRAS and BRAF; and microsatellite instability. Among the 19 CIMP-positive cancers, very few shared potential driver events, and those drivers were only IDH1 and SETD2 mutations. Finally, we found that CIMP was strongly correlated with tumor microenvironment characteristics, such as lymphocyte infiltration. Overall, our results indicate that CIMP does not exhibit a pan-cancer manifestation; rather, general dysregulation of CpG DNA methylation is caused by heterogeneous mechanisms.

Metagenomic next-generation sequencing (mNGS) enables comprehensive pathogen detection and has become increasingly popular in clinical diagnosis. The distinct pathogenic traits between strains require mNGS to achieve a strain-level resolution, but an equivocal concept of 'strain' as well as the low pathogen loads in most clinical specimens hinders such strain awareness. Here we introduce a metagenomic intra-species typing (MIST) tool (https://github.com/pandafengye/MIST), which hierarchically organizes reference genomes based on average nucleotide identity (ANI) and performs maximum likelihood estimation to infer the strain-level compositional abundance. In silico analysis using synthetic datasets showed that MIST accurately predicted the strain composition at a 99.9% average nucleotide identity (ANI) resolution with a merely 0.001× sequencing depth. When applying MIST on 359 culture-positive and 359 culture-negative real-world specimens of infected body fluids, we found the presence of multiple-strain reached considerable frequencies (30.39%-93.22%), which were otherwise underestimated by current diagnostic techniques due to their limited resolution. Several high-risk clones were identified to be prevalent across samples, including Acinetobacter baumannii sequence type (ST)208/ST195, Staphylococcus aureus ST22/ST398 and Klebsiella pneumoniae ST11/ST15, indicating potential outbreak events occurring in the clinical settings. Interestingly, contaminations caused by the engineered Escherichia coli strain K-12 and BL21 throughout the mNGS datasets were also identified by MIST instead of the statistical decontamination approach. Our study systemically characterized the infected body fluids at the strain level for the first time. Extension of mNGS testing to the strain level can greatly benefit clinical diagnosis of bacterial infections, including the identification of multi-strain infection, decontamination and infection control surveillance.

MOTIVATION: Effective computational methods to predict drug-protein interactions (DPIs) are vital for drug discovery in reducing the time and cost of drug development. Recent DPI prediction methods mainly exploit graph data composed of multiple kinds of connections among drugs and proteins. Each node in the graph usually has topological structures with multiple scales formed by its first-order neighbors and multi-order neighbors. However, most of the previous methods do not consider the topological structures of multi-order neighbors. In addition, deep integration of the multi-modality similarities of drugs and proteins is also a challenging task. RESULTS: We propose a model called ALDPI to adaptively learn the multi-scale topologies and multi-modality similarities with various significance levels. We first construct a drug-protein heterogeneous graph, which is composed of the interactions and the similarities with multiple modalities among drugs and proteins. An adaptive graph learning module is then designed to learn important kinds of connections in heterogeneous graph and generate new topology graphs. A module based on graph convolutional autoencoders is established to learn multiple representations, which imply the node attributes and multiple-scale topologies composed of one-order and multi-order neighbors, respectively. We also design an attention mechanism at neighbor topology level to distinguish the importance of these representations. Finally, since each similarity modality has its specific features, we construct a multi-layer convolutional neural network-based module to learn and fuse multi-modality features to obtain the attribute representation of each drug-protein node pair. Comprehensive experimental results show ALDPI's superior performance over six state-of-the-art methods. The results of recall rates of top-ranked candidates and case studies on five drugs further demonstrate the ability of ALDPI to discover potential drug-related protein candidates. CONTACT: zhang@hlju.edu.cn.

Protein turnover is vital for cellular functioning and is often associated with the pathophysiology of a variety of diseases. Metabolic labeling with heavy water followed by liquid chromatography coupled to mass spectrometry is a powerful tool to study in vivo protein turnover in high throughput and large scale. Heavy water is a cost-effective and easy to use labeling agent. It labels all nonessential amino acids. Due to its toxicity in high concentrations (20% or higher), small enrichments (8% or smaller) of heavy water are used with most organisms. The low concentration results in incomplete labeling of peptides/proteins. Therefore, the data processing is more challenging and requires accurate quantification of labeled and unlabeled forms of a peptide from overlapping mass isotopomer distributions. The work describes the bioinformatics aspects of the analysis of heavy water labeled mass spectral data, available software tools and current challenges and opportunities.

T-cell receptor (TCR) sequencing has enabled the development of innovative diagnostic tests for cancers, autoimmune diseases and other applications. However, the rarity of many T-cell clonotypes presents a detection challenge, which may lead to misdiagnosis if diagnostically relevant TCRs remain undetected. To address this issue, we developed TCRpower, a novel computational pipeline for quantifying the statistical detection power of TCR sequencing methods. TCRpower calculates the probability of detecting a TCR sequence as a function of several key parameters: in-vivo TCR frequency, T-cell sample count, read sequencing depth and read cutoff. To calibrate TCRpower, we selected unique TCRs of 45 T-cell clones (TCCs) as spike-in TCRs. We sequenced the spike-in TCRs from TCCs, together with TCRs from peripheral blood, using a 5' RACE protocol. The 45 spike-in TCRs covered a wide range of sample frequencies, ranging from 5 per 100 to 1 per 1 million. The resulting spike-in TCR read counts and ground truth frequencies allowed us to calibrate TCRpower. In our TCR sequencing data, we observed a consistent linear relationship between sample and sequencing read frequencies. We were also able to reliably detect spike-in TCRs with frequencies as low as one per million. By implementing an optimized read cutoff, we eliminated most of the falsely detected sequences in our data (TCR α-chain 99.0% and TCR β-chain 92.4%), thereby improving diagnostic specificity. TCRpower is publicly available and can be used to optimize future TCR sequencing experiments, and thereby enable reliable detection of disease-relevant TCRs for diagnostic applications.

Combination therapy has shown an obvious curative effect on complex diseases, whereas the search space of drug combinations is too large to be validated experimentally even with high-throughput screens. With the increase of the number of drugs, artificial intelligence techniques, especially machine learning methods, have become applicable for the discovery of synergistic drug combinations to significantly reduce the experimental workload. In this study, in order to predict novel synergistic drug combinations in various cancer cell lines, the cell line-specific drug-induced gene expression profile (GP) is added as a new feature type to capture the cellular response of drugs and reveal the biological mechanism of synergistic effect. Then, an enhanced cascade-based deep forest regressor (EC-DFR) is innovatively presented to apply the new small-scale drug combination dataset involving chemical, physical and biological (GP) properties of drugs and cells. Verified by the dataset, EC-DFR outperforms two state-of-the-art deep neural network-based methods and several advanced classical machine learning algorithms. Biological experimental validation performed subsequently on a set of previously untested drug combinations further confirms the performance of EC-DFR. What is more prominent is that EC-DFR can distinguish the most important features, making it more interpretable. By evaluating the contribution of each feature type, GP feature contributes 82.40%, showing the cellular responses of drugs may play crucial roles in synergism prediction. The analysis based on the top contributing genes in GP further demonstrates some potential relationships between the transcriptomic levels of key genes under drug regulation and the synergism of drug combinations.

Normalization and batch correction are critical steps in processing single-cell RNA sequencing (scRNA-seq) data, which remove technical effects and systematic biases to unmask biological signals of interest. Although a number of computational methods have been developed, there is no guidance for choosing appropriate procedures in different scenarios. In this study, we assessed the performance of 28 scRNA-seq noise reduction procedures in 55 scenarios using simulated and real datasets. The scenarios accounted for multiple biological and technical factors that greatly affect the denoising performance, including relative magnitude of batch effects, the extent of cell population imbalance, the complexity of cell group structures, the proportion and the similarity of nonoverlapping cell populations, dropout rates and variable library sizes. We used multiple quantitative metrics and visualization of low-dimensional cell embeddings to evaluate the performance on batch mixing while preserving the original cell group and gene structures. Based on our results, we specified technical or biological factors affecting the performance of each method and recommended proper methods in different scenarios. In addition, we highlighted one challenging scenario where most methods failed and resulted in overcorrection. Our studies not only provided a comprehensive guideline for selecting suitable noise reduction procedures but also pointed out unsolved issues in the field, especially the urgent need of developing metrics for assessing batch correction on imperceptible cell-type mixing.

The increasing volume of ribosome profiling (Ribo-seq) data, computational complexity of its data processing and operational handicap of related analytical procedures present a daunting set of informatics challenges. These impose a substantial barrier to researchers particularly with no or limited bioinformatics expertise in analyzing and decoding translation information from Ribo-seq data, thus driving the need for a new research paradigm for data computation and information extraction. In this knowledge base, we herein present a novel interactive web platform, RiboChat (https://db.cngb.org/ribobench/chat.html), for direct analyzing and annotating Ribo-seq data in the form of a chat conversation. It consists of a user-friendly web interface and a backend cloud-computing service. When typing a data analysis question into the chat window, the object-text detection module will be run to recognize relevant keywords from the input text. Based on the features identified in the input, individual analytics modules are then scored to find the perfect-matching candidate. The corresponding analytics module will be further executed after checking the completion status of the uploading of datasets and configured parameters. Overall, RiboChat represents an important step forward in the emerging direction of next-generation data analytics and will enable the broad research community to conveniently decipher translation information embedded within Ribo-seq data.

Advances in single-cell biotechnologies simultaneously generate the transcriptomic and epigenomic profiles at cell levels, providing an opportunity for investigating cell fates. Although great efforts have been devoted to either of them, the integrative analysis of single-cell multi-omics data is really limited because of the heterogeneity, noises and sparsity of single-cell profiles. In this study, a network-based integrative clustering algorithm (aka NIC) is present for the identification of cell types by fusing the parallel single-cell transcriptomic (scRNA-seq) and epigenomic profiles (scATAC-seq or DNA methylation). To avoid heterogeneity of multi-omics data, NIC automatically learns the cell-cell similarity graphs, which transforms the fusion of multi-omics data into the analysis of multiple networks. Then, NIC employs joint non-negative matrix factorization to learn the shared features of cells by exploiting the structure of learned cell-cell similarity networks, providing a better way to characterize the features of cells. The graph learning and integrative analysis procedures are jointly formulated as an optimization problem, and then the update rules are derived. Thirteen single-cell multi-omics datasets from various tissues and organisms are adopted to validate the performance of NIC, and the experimental results demonstrate that the proposed algorithm significantly outperforms the state-of-the-art methods in terms of various measurements. The proposed algorithm provides an effective strategy for the integrative analysis of single-cell multi-omics data (The software is coded using Matlab, and is freely available for academic https://github.com/xkmaxidian/NIC ).

Deep learning shortens the cycle of the drug discovery for its success in extracting features of molecules and proteins. Generating new molecules with deep learning methods could enlarge the molecule space and obtain molecules with specific properties. However, it is also a challenging task considering that the connections between atoms are constrained by chemical rules. Aiming at generating and optimizing new valid molecules, this article proposed Molecular Substructure Tree Generative Model, in which the molecule is generated by adding substructure gradually. The proposed model is based on the Variational Auto-Encoder architecture, which uses the encoder to map molecules to the latent vector space, and then builds an autoregressive generative model as a decoder to generate new molecules from Gaussian distribution. At the same time, for the molecular optimization task, a molecular optimization model based on CycleGAN was constructed. Experiments showed that the model could generate valid and novel molecules, and the optimized model effectively improves the molecular properties.

Drug repositioning is an efficient and promising strategy for traditional drug discovery and development. Many research efforts are focused on utilizing deep-learning approaches based on a heterogeneous network for modeling complex drug-disease associations. Similar to traditional latent factor models, which directly factorize drug-disease associations, they assume the neighbors are independent of each other in the network and thus tend to be ineffective to capture localized information. In this study, we propose a novel neighborhood and neighborhood interaction-based neural collaborative filtering approach (called DRWBNCF) to infer novel potential drugs for diseases. Specifically, we first construct three networks, including the known drug-disease association network, the drug-drug similarity and disease-disease similarity networks (using the nearest neighbors). To take the advantage of localized information in the three networks, we then design an integration component by proposing a new weighted bilinear graph convolution operation to integrate the information of the known drug-disease association, the drug's and disease's neighborhood and neighborhood interactions into a unified representation. Lastly, we introduce a prediction component, which utilizes the multi-layer perceptron optimized by the α-balanced focal loss function and graph regularization to model the complex drug-disease associations. Benchmarking comparisons on three datasets verified the effectiveness of DRWBNCF for drug repositioning. Importantly, the unknown drug-disease associations predicted by DRWBNCF were validated against clinical trials and three authoritative databases and we listed several new DRWBNCF-predicted potential drugs for breast cancer (e.g. valrubicin and teniposide) and small cell lung cancer (e.g. valrubicin and cytarabine).

Small proteins encoded by short open reading frames (ORFs) with 50 codons or fewer are emerging as an important class of cellular macromolecules in diverse organisms. However, they often evade detection by proteomics or in silico methods. Ribosome profiling (Ribo-seq) has revealed widespread translation in genomic regions previously thought to be non-coding, driving the development of ORF detection tools using Ribo-seq data. However, only a handful of tools have been designed for bacteria, and these have not yet been systematically compared. Here, we aimed to identify tools that use Ribo-seq data to correctly determine the translational status of annotated bacterial ORFs and also discover novel translated regions with high sensitivity. To this end, we generated a large set of annotated ORFs from four diverse bacterial organisms, manually labeled for their translation status based on Ribo-seq data, which are available for future benchmarking studies. This set was used to investigate the predictive performance of seven Ribo-seq-based ORF detection tools (REPARATION_blast, DeepRibo, Ribo-TISH, PRICE, smORFer, ribotricer and SPECtre), as well as IRSOM, which uses coding potential and RNA-seq coverage only. DeepRibo and REPARATION_blast robustly predicted translated ORFs, including sORFs, with no significant difference for ORFs in close proximity to other genes versus stand-alone genes. However, no tool predicted a set of novel, experimentally verified sORFs with high sensitivity. Start codon predictions with smORFer show the value of initiation site profiling data to further improve the sensitivity of ORF prediction tools in bacteria. Overall, we find that bacterial tools perform well for sORF detection, although there is potential for improving their performance, applicability, usability and reproducibility.

With the increasing volume of human sequencing data available, analysis incorporating external controls becomes a popular and cost-effective approach to boost statistical power in disease association studies. To prevent spurious association due to population stratification, it is important to match the ancestry backgrounds of cases and controls. However, rare variant association tests based on a standard logistic regression model are conservative when all ancestry-matched strata have the same case-control ratio and might become anti-conservative when case-control ratio varies across strata. Under the conditional logistic regression (CLR) model, we propose a weighted burden test (CLR-Burden), a variance component test (CLR-SKAT) and a hybrid test (CLR-MiST). We show that the CLR model coupled with ancestry matching is a general approach to control for population stratification, regardless of the spatial distribution of disease risks. Through extensive simulation studies, we demonstrate that the CLR-based tests robustly control type 1 errors under different matching schemes and are more powerful than the standard Burden, SKAT and MiST tests. Furthermore, because CLR-based tests allow for different case-control ratios across strata, a full-matching scheme can be employed to efficiently utilize all available cases and controls to accelerate the discovery of disease associated genes.

The development of autoimmune diseases following SARS-CoV-2 infection, including multisystem inflammatory syndrome, has been reported, and several mechanisms have been suggested, including molecular mimicry. We developed a scalable, comparative immunoinformatics pipeline called cross-reactive-epitope-search-using-structural-properties-of-proteins (CRESSP) to identify cross-reactive epitopes between a collection of SARS-CoV-2 proteomes and the human proteome using the structural properties of the proteins. Overall, by searching 4 911 245 proteins from 196 352 SARS-CoV-2 genomes, we identified 133 and 648 human proteins harboring potential cross-reactive B-cell and CD8+ T-cell epitopes, respectively. To demonstrate the robustness of our pipeline, we predicted the cross-reactive epitopes of coronavirus spike proteins, which were recognized by known cross-neutralizing antibodies. Using single-cell expression data, we identified PARP14 as a potential target of intermolecular epitope spreading between the virus and human proteins. Finally, we developed a web application (https://ahs2202.github.io/3M/) to interactively visualize our results. We also made our pipeline available as an open-source CRESSP package (https://pypi.org/project/cressp/), which can analyze any two proteomes of interest to identify potentially cross-reactive epitopes between the proteomes. Overall, our immunoinformatic resources provide a foundation for the investigation of molecular mimicry in the pathogenesis of autoimmune and chronic inflammatory diseases following COVID-19.

Microbiome research is advancing rapidly, and every new study should definitively be based on updated methods, trends and milestones in this field to avoid the wrong interpretation of results. Most human microbiota surveys rely on data captured from snapshots-single data points from subjects-and have permitted uncovering the recognized interindividual variability and major covariates of such microbial communities. Currently, changes in individualized microbiota profiles are under the spotlight to serve as robust predictors of clinical outcomes (e.g. weight loss via dietary interventions) and disease anticipation. Therefore, novel methods are needed to provide robust evaluation of longitudinal series of microbiota data with the aim of assessing intrapersonally short-term to long-term microbiota changes likely linked to health and disease states. Consequently, we developed microbiota STability ASsessment via Iterative cluStering (μSTASIS)-a multifunction R package to evaluate individual-centered microbiota stability. μSTASIS targets the recognized interindividual variability inherent to microbiota data to stress the tight relationships observed among and characteristic of longitudinal samples derived from a single individual via iteratively growing-partitioned clustering. The algorithms and functions implemented in this framework deal properly with the sparse and compositional nature of microbiota data. Moreover, the resulting metric is intuitive and independent of beta diversity distance methods and correlation coefficients, thus estimating stability for each microbiota sample rather than giving nonconsensus magnitudes that are difficult to interpret within and between datasets. Our method is freely available under GPL-3 licensing. We demonstrate its utility by assessing gut microbiota stability from three independent studies published previously with multiple longitudinal series of multivariate data and respective metadata.

N6-methyladenine (6mA) is associated with important roles in DNA replication, DNA repair, transcription, regulation of gene expression. Several experimental methods were used to identify DNA modifications. However, these experimental methods are costly and time-consuming. To detect the 6mA and complement these shortcomings of experimental methods, we proposed a novel, deep leaning approach called BERT6mA. To compare the BERT6mA with other deep learning approaches, we used the benchmark datasets including 11 species. The BERT6mA presented the highest AUCs in eight species in independent tests. Furthermore, BERT6mA showed higher and comparable performance with the state-of-the-art models while the BERT6mA showed poor performances in a few species with a small sample size. To overcome this issue, pretraining and fine-tuning between two species were applied to the BERT6mA. The pretrained and fine-tuned models on specific species presented higher performances than other models even for the species with a small sample size. In addition to the prediction, we analyzed the attention weights generated by BERT6mA to reveal how the BERT6mA model extracts critical features responsible for the 6mA prediction. To facilitate biological sciences, the BERT6mA online web server and its source codes are freely accessible at https://github.com/kuratahiroyuki/BERT6mA.git, respectively.

A major topic of debate in developmental biology centers on whether development is continuous, discontinuous, or a mixture of both. Pseudo-time trajectory models, optimal for visualizing cellular progression, model cell transitions as continuous state manifolds and do not explicitly model real-time, complex, heterogeneous systems and are challenging for benchmarking with temporal models. We present a data-driven framework that addresses these limitations with temporal single-cell data collected at discrete time points as inputs and a mixture of dependent minimum spanning trees (MSTs) as outputs, denoted as dynamic spanning forest mixtures (DSFMix). DSFMix uses decision-tree models to select genes that account for variations in multimodality, skewness and time. The genes are subsequently used to build the forest using tree agglomerative hierarchical clustering and dynamic branch cutting. We first motivate the use of forest-based algorithms compared to single-tree approaches for visualizing and characterizing developmental processes. We next benchmark DSFMix to pseudo-time and temporal approaches in terms of feature selection, time correlation, and network similarity. Finally, we demonstrate how DSFMix can be used to visualize, compare and characterize complex relationships during biological processes such as epithelial-mesenchymal transition, spermatogenesis, stem cell pluripotency, early transcriptional response from hormones and immune response to coronavirus disease. Our results indicate that the expression of genes during normal development exhibits a high proportion of non-uniformly distributed profiles that are mostly right-skewed and multimodal; the latter being a characteristic of major steady states during development. Our study also identifies and validates gene signatures driving complex dynamic processes during somatic or germline differentiation.

Protein-protein interactions (PPIs) play a significant role in nearly all cellular and biological activities. Data-driven machine learning models have demonstrated great power in PPIs. However, the design of efficient molecular featurization poses a great challenge for all learning models for PPIs. Here, we propose persistent spectral (PerSpect) based PPI representation and featurization, and PerSpect-based ensemble learning (PerSpect-EL) models for PPI binding affinity prediction, for the first time. In our model, a sequence of Hodge (or combinatorial) Laplacian (HL) matrices at various different scales are generated from a specially designed filtration process. PerSpect attributes, which are statistical and combinatorial properties of spectrum information from these HL matrices, are used as features for PPI characterization. Each PerSpect attribute is input into a 1D convolutional neural network (CNN), and these CNN networks are stacked together in our PerSpect-based ensemble learning models. We systematically test our model on the two most commonly used datasets, i.e. SKEMPI and AB-Bind. It has been found that our model can achieve state-of-the-art results and outperform all existing models to the best of our knowledge.

Protein lysine crotonylation (Kcr) is an important type of posttranslational modification that is associated with a wide range of biological processes. The identification of Kcr sites is critical to better understanding their functional mechanisms. However, the existing experimental techniques for detecting Kcr sites are cost-ineffective, to a great need for new computational methods to address this problem. We here describe Adapt-Kcr, an advanced deep learning model that utilizes adaptive embedding and is based on a convolutional neural network together with a bidirectional long short-term memory network and attention architecture. On the independent testing set, Adapt-Kcr outperformed the current state-of-the-art Kcr prediction model, with an improvement of 3.2% in accuracy and 1.9% in the area under the receiver operating characteristic curve. Compared to other Kcr models, Adapt-Kcr additionally had a more robust ability to distinguish between crotonylation and other lysine modifications. Another model (Adapt-ST) was trained to predict phosphorylation sites in SARS-CoV-2, and outperformed the equivalent state-of-the-art phosphorylation site prediction model. These results indicate that self-adaptive embedding features perform better than handcrafted features in capturing discriminative information; when used in attention architecture, this could be an effective way of identifying protein Kcr sites. Together, our Adapt framework (including learning embedding features and attention architecture) has a strong potential for prediction of other protein posttranslational modification sites.

Subcellular localization of microRNAs (miRNAs) is an important reflection of their biological functions. Considering the spatio-temporal specificity of miRNA subcellular localization, experimental detection techniques are expensive and time-consuming, which strongly motivates an efficient and economical computational method to predict miRNA subcellular localization. In this paper, we describe a computational framework, MiRLoc, to predict the subcellular localization of miRNAs. In contrast to existing methods, MiRLoc uses the functional similarity between miRNAs instead of sequence features and incorporates information about the subcellular localization of the corresponding target mRNAs. The results show that miRNA functional similarity data can be effectively used to predict miRNA subcellular localization, and that inclusion of subcellular localization information of target mRNAs greatly improves prediction performance.

Chromosome is composed of many distinct chromatin domains, referred to variably as topological domains or topologically associating domains (TADs). The domains are stable across different cell types and highly conserved across species, thus these chromatin domains have been considered as the basic units of chromosome folding and regarded as an important secondary structure in chromosome organization. However, the identification of TAD boundaries is still a great challenge due to the high cost and low resolution of Hi-C data or experiments. In this study, we propose a novel ensemble learning framework, termed as StackTADB, for predicting the boundaries of TADs. StackTADB integrates four base classifiers including Random Forest, Logistic Regression, K-NearestNeighbor and Support Vector Machine. From the analysis of a series of examinations on the data set in the previous study, it is concluded that StackTADB has optimal performance in six metrics, AUC, Accuracy, MCC, Precision, Recall and F1 score, and it is superior to the existing methods. In addition, the comparison of the performance of multiple features shows that Kmers-based features play an essential role in predicting TADs boundaries of fruit flies, and we also apply the SHapley Additive exPlanations (SHAP) framework to interpret the predictions of StackTADB to identify the reason why Kmers-based features are vital. The experimental results show that the subsequences matching the BEAF-32 motif play a crucial role in predicting the boundaries of TADs. The source code is freely available at https://github.com/HaoWuLab-Bioinformatics/StackTADB and the webserver of StackTADB is freely available at http://hwtad.sdu.edu.cn:8002/StackTADB.

Single-cell RNA sequencing (scRNA-seq) permits researchers to study the complex mechanisms of cell heterogeneity and diversity. Unsupervised clustering is of central importance for the analysis of the scRNA-seq data, as it can be used to identify putative cell types. However, due to noise impacts, high dimensionality and pervasive dropout events, clustering analysis of scRNA-seq data remains a computational challenge. Here, we propose a new deep structural clustering method for scRNA-seq data, named scDSC, which integrate the structural information into deep clustering of single cells. The proposed scDSC consists of a Zero-Inflated Negative Binomial (ZINB) model-based autoencoder, a graph neural network (GNN) module and a mutual-supervised module. To learn the data representation from the sparse and zero-inflated scRNA-seq data, we add a ZINB model to the basic autoencoder. The GNN module is introduced to capture the structural information among cells. By joining the ZINB-based autoencoder with the GNN module, the model transfers the data representation learned by autoencoder to the corresponding GNN layer. Furthermore, we adopt a mutual supervised strategy to unify these two different deep neural architectures and to guide the clustering task. Extensive experimental results on six real scRNA-seq datasets demonstrate that scDSC outperforms state-of-the-art methods in terms of clustering accuracy and scalability. Our method scDSC is implemented in Python using the Pytorch machine-learning library, and it is freely available at https://github.com/DHUDBlab/scDSC.

Abnormal protein-membrane attachment is involved in deregulated cellular pathways and in disease. Therefore, the possibility to modulate protein-membrane interactions represents a new promising therapeutic strategy for peripheral membrane proteins that have been considered so far undruggable. A major obstacle in this drug design strategy is that the membrane-binding domains of peripheral membrane proteins are usually unknown. The development of fast and efficient algorithms predicting the protein-membrane interface would shed light into the accessibility of membrane-protein interfaces by drug-like molecules. Herein, we describe an ensemble machine learning methodology and algorithm for predicting membrane-penetrating amino acids. We utilize available experimental data from the literature for training 21 machine learning classifiers and meta-classifiers. Evaluation of the best ensemble classifier model accuracy yields a macro-averaged F1 score = 0.92 and a Matthews correlation coefficient = 0.84 for predicting correctly membrane-penetrating amino acids on unknown proteins of a validation set. The python code for predicting protein-membrane interfaces of peripheral membrane proteins is available at https://github.com/zoecournia/DREAMM.

Individual pathway analysis can dissect heterogeneities among different cancer patients and provide efficient guidelines for individualized therapy. However, the existence of the batch effect brings extensive limitations for the application of many individual methods for pathway analysis. Previously, researchers proposed that methods based on within-sample relative expression ordering (REO) of the genes are notably insensitive to 'batch effects'. In this article, we focus on the Gene Ontology (GO) database and propose an individual qualitative GO term analysis method (IndGOterm) based on the REO of genes. Compared with some current widely used single-sample enrichment analysis methods, such as ssGSEA and GSVA, IndGOterm has a predominance of ignoring the batch effects caused by diverse technologies. Through the survival and drug responses analysis, we found IndGOterm could capture more terms connected to cancer than other single-sample enrichment analysis methods. Furthermore, through the application of IndGOterm, we found some terms that present different dysregulation models that manifest heterogenetic in homologous patients. Collectively, these results attested that IndGOterm could capture useful information from patients and be a useful tool to reveal the intrinsic characteristic of cancer. An open-source R statistical analysis package 'IndGOterm' is available at https://github.com/robert19960424/IndGOterm.

Identifying differential genes over conditions provides insights into the mechanisms of biological processes and disease progression. Here we present an approach, the Kullback-Leibler divergence-based differential distribution (klDD), which provides a flexible framework for quantifying changes in higher-order statistical information of genes including mean and variance/covariation. The method can well detect subtle differences in gene expression distributions in contrast to mean or variance shifts of the existing methods. In addition to effectively identifying informational genes in terms of differential distribution, klDD can be directly applied to cancer subtyping, single-cell clustering and disease early-warning detection, which were all validated by various benchmark datasets.

The advances in single-cell RNA sequencing (scRNA-seq) technologies enable the characterization of transcriptomic profiles at the cellular level and demonstrate great promise in bulk sample analysis thereby offering opportunities to transfer gene signature from scRNA-seq to bulk data. However, the gene expression signatures identified from single cells are typically inapplicable to bulk RNA-seq data due to the profiling differences of distinct sequencing technologies. Here, we propose single-cell pair-wise gene expression (scPAGE), a novel method to develop single-cell gene pair signatures (scGPSs) that were beneficial to bulk RNA-seq classification to transfer knowledge across platforms. PAGE was adopted to tackle the challenge of profiling differences. We applied the method to acute myeloid leukemia (AML) and identified the scGPS from mouse scRNA-seq that allowed discriminating between AML and control cells. The scGPS was validated in bulk RNA-seq datasets and demonstrated better performance (average area under the curve [AUC] = 0.96) than the conventional gene expression strategies (average AUC$\le$ 0.88) suggesting its potential in disclosing the molecular mechanism of AML. The scGPS also outperformed its bulk counterpart, which highlighted the benefit of gene signature transfer. Furthermore, we confirmed the utility of scPAGE in sepsis as an example of other disease scenarios. scPAGE leveraged the advantages of single-cell profiles to enhance the analysis of bulk samples revealing great potential of transferring knowledge from single-cell to bulk transcriptome studies.

With advances in library construction protocols and next-generation sequencing technologies, viral metagenomic sequencing has become the major source for novel virus discovery. Conducting taxonomic classification for metagenomic data is an important means to characterize the viral composition in the underlying samples. However, RNA viruses are abundant and highly diverse, jeopardizing the sensitivity of comparison-based classification methods. To improve the sensitivity of read-level taxonomic classification, we developed an RNA-dependent RNA polymerase (RdRp) gene-based read classification tool RdRpBin. It combines alignment-based strategy with machine learning models in order to fully exploit the sequence properties of RdRp. We tested our method and compared its performance with the state-of-the-art tools on the simulated and real sequencing data. RdRpBin competes favorably with all. In particular, when the query RNA viruses share low sequence similarity with the known viruses ($\sim 0.4$), our tool can still maintain a higher F-score than the state-of-the-art tools. The experimental results on real data also showed that RdRpBin can classify more RNA viral reads with a relatively low false-positive rate. Thus, RdRpBin can be utilized to classify novel and diverged RNA viruses.

The gene ontology (GO) provides a hierarchical structure with a controlled vocabulary composed of terms describing functions and localization of gene products. Recent works propose vector representations, also known as embeddings, of GO terms that capture meaningful information about them. Significant performance improvements have been observed when these representations are used on diverse downstream tasks, such as the measurement of semantic similarity between GO terms and functional similarity between proteins. Despite the success shown by these approaches, existing embeddings of GO terms still fail to capture crucial structural features of the GO. Here, we present anc2vec, a novel protocol based on neural networks for constructing vector representations of GO terms by preserving three important ontological features: its ontological uniqueness, ancestors hierarchy and sub-ontology membership. The advantages of using anc2vec are demonstrated by systematic experiments on diverse tasks: visualization, sub-ontology prediction, inference of structurally related terms, retrieval of terms from aggregated embeddings, and prediction of protein-protein interactions. In these tasks, experimental results show that the performance of anc2vec representations is better than those of recent approaches. This demonstrates that higher performances on diverse tasks can be achieved by embeddings when the structure of the GO is better represented. Full source code and data are available at https://github.com/sinc-lab/anc2vec.

The inference of gene co-expression associations is one of the fundamental tasks for large-scale transcriptomic data analysis. Due to the high dimensionality and high noises in transcriptomic data, it is difficult to infer stable gene co-expression associations from single dataset. Meta-analysis of multisource data can effectively tackle this problem. We proposed Joint Embedding of multiple BIpartite Networks (JEBIN) to learn the low-dimensional consensus representation for genes by integrating multiple expression datasets. JEBIN infers gene co-expression associations in a nonlinear and global similarity manner and can integrate datasets with different distributions in linear time complexity with the gene and total sample size. The effectiveness and scalability of JEBIN were verified by simulation experiments, and its superiority over the commonly used integration methods was proved by three indexes on real biological datasets. Then, JEBIN was applied to study the gene co-expression patterns of hepatocellular carcinoma (HCC) based on multiple expression datasets of HCC and adjacent normal tissues, and further on latest HCC single-cell RNA-seq data. Results show that gene co-expressions are highly different between bulk and single-cell datasets. Finally, many differentially co-expressed ligand-receptor pairs were discovered by comparing HCC with adjacent normal data, providing candidate HCC targets for abnormal cell-cell communications.

Feature representation and discriminative learning are proven models and technologies in artificial intelligence fields; however, major challenges for machine learning on large biological datasets are learning an effective model with mechanistical explanation on the model determination and prediction. To satisfy such demands, we developed Vec2image, an explainable convolutional neural network framework for characterizing the feature engineering, feature selection and classifier training that is mainly based on the collaboration of principal component coordinate conversion, deep residual neural networks and embedded k-nearest neighbor representation on pseudo images of high-dimensional biological data, where the pseudo images represent feature measurements and feature associations simultaneously. Vec2image has achieved better performance compared with other popular methods and illustrated its efficiency on feature selection in cell marker identification from tissue-specific single-cell datasets. In particular, in a case study on type 2 diabetes (T2D) by multiple human islet scRNA-seq datasets, Vec2image first displayed robust performance on T2D classification model building across different datasets, then a specific Vec2image model was trained to accurately recognize the cell state and efficiently rank feature genes relevant to T2D which uncovered potential T2D cellular pathogenesis; and next the cell activity changes, cell composition imbalances and cell-cell communication dysfunctions were associated to our finding T2D feature genes from both population-shared and individual-specific perspectives. Collectively, Vec2image is a new and efficient explainable artificial intelligence methodology that can be widely applied in human-readable classification and prediction on the basis of pseudo image representation of biological deep sequencing data.

Inferring gene regulatory networks (GRNs) based on gene expression profiles is able to provide an insight into a number of cellular phenotypes from the genomic level and reveal the essential laws underlying various life phenomena. Different from the bulk expression data, single-cell transcriptomic data embody cell-to-cell variance and diverse biological information, such as tissue characteristics, transformation of cell types, etc. Inferring GRNs based on such data offers unprecedented advantages for making a profound study of cell phenotypes, revealing gene functions and exploring potential interactions. However, the high sparsity, noise and dropout events of single-cell transcriptomic data pose new challenges for regulation identification. We develop a hybrid deep learning framework for GRN inference from single-cell transcriptomic data, DGRNS, which encodes the raw data and fuses recurrent neural network and convolutional neural network (CNN) to train a model capable of distinguishing related gene pairs from unrelated gene pairs. To overcome the limitations of such datasets, it applies sliding windows to extract valuable features while preserving the direction of regulation. DGRNS is constructed as a deep learning model containing gated recurrent unit network for exploring time-dependent information and CNN for learning spatially related information. Our comprehensive and detailed comparative analysis on the dataset of mouse hematopoietic stem cells illustrates that DGRNS outperforms state-of-the-art methods. The networks inferred by DGRNS are about 16% higher than the area under the receiver operating characteristic curve of other unsupervised methods and 10% higher than the area under the precision recall curve of other supervised methods. Experiments on human datasets show the strong robustness and excellent generalization of DGRNS. By comparing the predictions with standard network, we discover a series of novel interactions which are proved to be true in some specific cell types. Importantly, DGRNS identifies a series of regulatory relationships with high confidence and functional consistency, which have not yet been experimentally confirmed and merit further research.

Enhancer-promoter interaction (EPI) is a key mechanism underlying gene regulation. EPI prediction has always been a challenging task because enhancers could regulate promoters of distant target genes. Although many machine learning models have been developed, they leverage only the features in enhancers and promoters, or simply add the average genomic signals in the regions between enhancers and promoters, without utilizing detailed features between or outside enhancers and promoters. Due to a lack of large-scale features, existing methods could achieve only moderate performance, especially for predicting EPIs in different cell types. Here, we present a Transformer-based model, TransEPI, for EPI prediction by capturing large genomic contexts. TransEPI was developed based on EPI datasets derived from Hi-C or ChIA-PET data in six cell lines. To avoid over-fitting, we evaluated the TransEPI model by testing it on independent test datasets where the cell line and chromosome are different from the training data. TransEPI not only achieved consistent performance across the cross-validation and test datasets from different cell types but also outperformed the state-of-the-art machine learning and deep learning models. In addition, we found that the improved performance of TransEPI was attributed to the integration of large genomic contexts. Lastly, TransEPI was extended to study the non-coding mutations associated with brain disorders or neural diseases, and we found that TransEPI was also useful for predicting the target genes of non-coding mutations.

Accurate prediction of atomic partial charges with high-level quantum mechanics (QM) methods suffers from high computational cost. Numerous feature-engineered machine learning (ML)-based predictors with favorable computability and reliability have been developed as alternatives. However, extensive expertise effort was needed for feature engineering of atom chemical environment, which may consequently introduce domain bias. In this study, SuperAtomicCharge, a data-driven deep graph learning framework, was proposed to predict three important types of partial charges (i.e. RESP, DDEC4 and DDEC78) derived from high-level QM calculations based on the structures of molecules. SuperAtomicCharge was designed to simultaneously exploit the 2D and 3D structural information of molecules, which was proved to be an effective way to improve the prediction accuracy of the model. Moreover, a simple transfer learning strategy and a multitask learning strategy based on self-supervised descriptors were also employed to further improve the prediction accuracy of the proposed model. Compared with the latest baselines, including one GNN-based predictor and two ML-based predictors, SuperAtomicCharge showed better performance on all the three external test sets and had better usability and portability. Furthermore, the QM partial charges of new molecules predicted by SuperAtomicCharge can be efficiently used in drug design applications such as structure-based virtual screening, where the predicted RESP and DDEC4 charges of new molecules showed more robust scoring and screening power than the commonly used partial charges. Finally, two tools including an online server (http://cadd.zju.edu.cn/deepchargepredictor) and the source code command lines (https://github.com/zjujdj/SuperAtomicCharge) were developed for the easy access of the SuperAtomicCharge services.

Genomic epidemiology is important to study the COVID-19 pandemic, and more than two million severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) genomic sequences were deposited into public databases. However, the exponential increase of sequences invokes unprecedented bioinformatic challenges. Here, we present the Coronavirus GenBrowser (CGB) based on a highly efficient analysis framework and a node-picking rendering strategy. In total, 1,002,739 high-quality genomic sequences with the transmission-related metadata were analyzed and visualized. The size of the core data file is only 12.20 MB, highly efficient for clean data sharing. Quick visualization modules and rich interactive operations are provided to explore the annotated SARS-CoV-2 evolutionary tree. CGB binary nomenclature is proposed to name each internal lineage. The pre-analyzed data can be filtered out according to the user-defined criteria to explore the transmission of SARS-CoV-2. Different evolutionary analyses can also be easily performed, such as the detection of accelerated evolution and ongoing positive selection. Moreover, the 75 genomic spots conserved in SARS-CoV-2 but non-conserved in other coronaviruses were identified, which may indicate the functional elements specifically important for SARS-CoV-2. The CGB was written in Java and JavaScript. It not only enables users who have no programming skills to analyze millions of genomic sequences, but also offers a panoramic vision of the transmission and evolution of SARS-CoV-2.

Accurate simulation of protein folding is a unique challenge in understanding the physical process of protein folding, with important implications for protein design and drug discovery. Molecular dynamics simulation strongly requires advanced force fields with high accuracy to achieve correct folding. However, the current force fields are inaccurate, inapplicable and inefficient. We propose a machine learning protocol, the inductive transfer learning force field (ITLFF), to construct protein force fields in seconds with any level of accuracy from a small dataset. This process is achieved by incorporating an inductive transfer learning algorithm into deep neural networks, which learn knowledge of any high-level calculations from a large dataset of low-level method. Here, we use a double-hybrid density functional theory (DFT) as a case functional, but ITLFF is suitable for any high-precision functional. The performance of the selected 18 proteins indicates that compared with the fragment-based double-hybrid DFT algorithm, the force field constructed by ITLFF achieves considerable accuracy with a mean absolute error of 0.0039 kcal/mol/atom for energy and a root mean square error of 2.57 $\mathrm{kcal}/\mathrm{mol}/{\AA}$ for force, and it is more than 30 000 times faster and obtains more significant efficiency benefits as the system increases. The outstanding performance of ITLFF provides promising prospects for accurate and efficient protein dynamic simulations and makes an important step toward protein folding simulation. Due to the ability of ITLFF to utilize the knowledge acquired in one task to solve related problems, it is also applicable for various problems in biology, chemistry and material science.

There is a lack of robust generalizable predictive biomarkers of response to immune checkpoint blockade in multiple types of cancer. We develop hDirect-MAP, an algorithm that maps T cells into a shared high-dimensional (HD) expression space of diverse T cell functional signatures in which cells group by the common T cell phenotypes rather than dimensional reduced features or a distorted view of these features. Using projection-free single-cell modeling, hDirect-MAP first removed a large group of cells that did not contribute to response and then clearly distinguished T cells into response-specific subpopulations that were defined by critical T cell functional markers of strong differential expression patterns. We found that these grouped cells cannot be distinguished by dimensional-reduction algorithms but are blended by diluted expression patterns. Moreover, these identified response-specific T cell subpopulations enabled a generalizable prediction by their HD metrics. Tested using five single-cell RNA-seq or mass cytometry datasets from basal cell carcinoma, squamous cell carcinoma and melanoma, hDirect-MAP demonstrated common response-specific T cell phenotypes that defined a generalizable and accurate predictive biomarker.

Health outcomes are frequently shaped by difficult to dissect inter-relationships between biological, behavioral, social and environmental factors. DNA methylation patterns reflect such multivariate intersections, providing a rich source of novel biomarkers and insight into disease etiologies. Recent advances in whole-genome bisulfite sequencing enable investigation of DNA methylation over all genomic CpGs, but existing bioinformatic approaches lack accessible system-level tools. Here, we develop the R package Comethyl, for weighted gene correlation network analysis of user-defined genomic regions that generates modules of comethylated regions, which are then tested for correlations with multivariate sample traits. First, regions are defined by CpG genomic location or regulatory annotation and filtered based on CpG count, sequencing depth and variability. Next, correlation networks are used to find modules of interconnected nodes using methylation values within the selected regions. Each module containing multiple comethylated regions is reduced in complexity to a single eigennode value, which is then tested for correlations with experimental metadata. Comethyl has the ability to cover the noncoding regulatory regions of the genome with high relevance to interpretation of genome-wide association studies and integration with other types of epigenomic data. We demonstrate the utility of Comethyl on a dataset of male cord blood samples from newborns later diagnosed with autism spectrum disorder (ASD) versus typical development. Comethyl successfully identified an ASD-associated module containing regions mapped to genes enriched for brain glial functions. Comethyl is expected to be useful in uncovering the multivariate nature of health disparities for a variety of common disorders. Comethyl is available at github.com/cemordaunt/comethyl with complete documentation and example analyses.

Networks consisting of molecular interactions are intrinsically dynamical systems of an organism. These interactions curated in molecular interaction databases are still not complete and contain false positives introduced by high-throughput screening experiments. In this study, we propose a framework to integrate interactions of functional associated protein-coding genes from 31 data sources to reconstruct a network with high coverage and quality. For each interaction, 369 features were constructed including properties of both the interaction and the involved genes. The training and validation sets were built on the pathway interactions as positives and the potential negative instances resulting from our proposed semi-supervised strategy. Random forest classification method was then applied to train and predict multiple times to give a score for each interaction. After setting a threshold estimated by a Binomial distribution, a Human protein-coding Gene Functional Association Network (HuGFAN) was reconstructed with 20 383 genes and 1185 429 high confidence interactions. Then, HuGFAN was compared with other networks from data sources with respect to network properties, suggesting that HuGFAN is more function and pathway related. Finally, HuGFAN was applied to identify cancer driver through two famous network-based methods (DriverNet and HotNet2) to show its outstanding performance compared with other networks. HuGFAN and other supplementary files are freely available at https://github.com/xthuang226/HuGFAN.

Single-cell RNA sequencing (scRNA-seq) techniques provide high-resolution data on cellular heterogeneity in diverse tissues, and a critical step for the data analysis is cell type identification. Traditional methods usually cluster the cells and manually identify cell clusters through marker genes, which is time-consuming and subjective. With the launch of several large-scale single-cell projects, millions of sequenced cells have been annotated and it is promising to transfer labels from the annotated datasets to newly generated datasets. One powerful way for the transferring is to learn cell relations through the graph neural network (GNN), but traditional GNNs are difficult to process millions of cells due to the expensive costs of the message-passing procedure at each training epoch. Here, we have developed a robust and scalable GNN-based method for accurate single-cell classification (GraphCS), where the graph is constructed to connect similar cells within and between labelled and unlabeled scRNA-seq datasets for propagation of shared information. To overcome the slow information propagation of GNN at each training epoch, the diffused information is pre-calculated via the approximate Generalized PageRank algorithm, enabling sublinear complexity over cell numbers. Compared with existing methods, GraphCS demonstrates better performance on simulated, cross-platform, cross-species and cross-omics scRNA-seq datasets. More importantly, our model provides a high speed and scalability on large datasets, and can achieve superior performance for 1 million cells within 50 min.

Comment on Brief Bioinform. 2021 Sep 2;22(5):

Accurate identification of drug-target interactions (DTIs) plays a crucial role in drug discovery. Compared with traditional experimental methods that are labor-intensive and time-consuming, computational methods are more and more popular in recent years. Conventional computational methods almost simply view heterogeneous networks which integrate diverse drug-related and target-related dataset instead of fully exploring drug and target similarities. In this paper, we propose a new method, named DTIHNC, for $\mathbf{D}$rug-$\mathbf{T}$arget $\mathbf{I}$nteraction identification, which integrates $\mathbf{H}$eterogeneous $\mathbf{N}$etworks and $\mathbf{C}$ross-modal similarities calculated by relations between drugs, proteins, diseases and side effects. Firstly, the low-dimensional features of drugs, proteins, diseases and side effects are obtained from original features by a denoising autoencoder. Then, we construct a heterogeneous network across drug, protein, disease and side-effect nodes. In heterogeneous network, we exploit the heterogeneous graph attention operations to update the embedding of a node based on information in its 1-hop neighbors, and for multi-hop neighbor information, we propose random walk with restart aware graph attention to integrate more information through a larger neighborhood region. Next, we calculate cross-modal drug and protein similarities from cross-scale relations between drugs, proteins, diseases and side effects. Finally, a multiple-layer convolutional neural network deeply integrates similarity information of drugs and proteins with the embedding features obtained from heterogeneous graph attention network. Experiments have demonstrated its effectiveness and better performance than state-of-the-art methods. Datasets and a stand-alone package are provided on Github with website https://github.com/ningq669/DTIHNC.

Although sifting functional genes has been discussed for years, traditional selection methods tend to be ineffective in capturing potential specific genes. First, typical methods focus on finding features (genes) relevant to class while irrelevant to each other. However, the features that can offer rich discriminative information are more likely to be the complementary ones. Next, almost all existing methods assess feature relations in pairs, yielding an inaccurate local estimation and lacking a global exploration. In this paper, we introduce multi-variable Area Under the receiver operating characteristic Curve (AUC) to globally evaluate the complementarity among features by employing Area Above the receiver operating characteristic Curve (AAC). Due to AAC, the class-relevant information newly provided by a candidate feature and that preserved by the selected features can be achieved beyond pairwise computation. Furthermore, we propose an AAC-based feature selection algorithm, named Multi-variable AUC-based Combined Features Complementarity, to screen discriminative complementary feature combinations. Extensive experiments on public datasets demonstrate the effectiveness of the proposed approach. Besides, we provide a gene set about prostate cancer and discuss its potential biological significance from the machine learning aspect and based on the existing biomedical findings of some individual genes.

Integration of expression quantitative trait loci (eQTL) into genome-wide association studies (GWASs) is a promising manner to reveal functional roles of associated single-nucleotide polymorphisms (SNPs) in complex phenotypes and has become an active research field in post-GWAS era. However, how to efficiently incorporate eQTL mapping study into GWAS for prioritization of causal genes remains elusive. We herein proposed a novel method termed as Mixed transcriptome-wide association studies (TWAS) and mediated Variance estimation (MTV) by modeling the effects of cis-SNPs of a gene as a function of eQTL. MTV formulates the integrative method and TWAS within a unified framework via mixed models and therefore includes many prior methods/tests as special cases. We further justified MTV from another two statistical perspectives of mediation analysis and two-stage Mendelian randomization. Relative to existing methods, MTV is superior for pronounced features including the processing of direct effects of cis-SNPs on phenotypes, the powerful likelihood ratio test for assessment of joint effects of cis-SNPs and genetically regulated gene expression (GReX), two useful quantities to measure relative genetic contributions of GReX and cis-SNPs to phenotypic variance, and the computationally efferent parameter expansion expectation maximum algorithm. With extensive simulations, we identified that MTV correctly controlled the type I error in joint evaluation of the total genetic effect and proved more powerful to discover true association signals across various scenarios compared to existing methods. We finally applied MTV to 41 complex traits/diseases available from three GWASs and discovered many new associated genes that had otherwise been missed by existing methods. We also revealed that a small but substantial fraction of phenotypic variation was mediated by GReX. Overall, MTV constructs a robust and realistic modeling foundation for integrative omics analysis and has the advantage of offering more attractive biological interpretations of GWAS results.

Antimicrobial resistance (AMR) is a growing threat to public health and farming at large. In clinical and veterinary practice, timely characterization of the antibiotic susceptibility profile of bacterial infections is a crucial step in optimizing treatment. High-throughput sequencing is a promising option for clinical point-of-care and ecological surveillance, opening the opportunity to develop genotyping-based AMR determination as a possibly faster alternative to phenotypic testing. In the present work, we compare the performance of state-of-the-art methods for detection of AMR using high-throughput sequencing data from clinical settings. We consider five computational approaches based on alignment (AMRPlusPlus), deep learning (DeepARG), k-mer genomic signatures (KARGA, ResFinder) or hidden Markov models (Meta-MARC). We use an extensive collection of 585 isolates with available AMR resistance profiles determined by phenotypic tests across nine antibiotic classes. We show how the prediction landscape of AMR classifiers is highly heterogeneous, with balanced accuracy varying from 0.40 to 0.92. Although some algorithms-ResFinder, KARGA and AMRPlusPlus-exhibit overall better balanced accuracy than others, the high per-AMR-class variance and related findings suggest that: (1) all algorithms might be subject to sampling bias both in data repositories used for training and experimental/clinical settings; and (2) a portion of clinical samples might contain uncharacterized AMR genes that the algorithms-mostly trained on known AMR genes-fail to generalize upon. These results lead us to formulate practical advice for software configuration and application, and give suggestions for future study designs to further develop AMR prediction tools from proof-of-concept to bedside.

Herbicides have revolutionised weed management, increased crop yields and improved profitability allowing for an increase in worldwide food security. Their widespread use, however, has also led to a rise in resistance and concerns about their environmental impact. Despite the need for potent and safe herbicidal molecules, no herbicide with a new mode of action has reached the market in 30 years. Although development of computational approaches has proven invaluable to guide rational drug discovery pipelines, leading to higher hit rates and lower attrition due to poor toxicity, little has been done in contrast for herbicide design. To fill this gap, we have developed cropCSM, a computational platform to help identify new, potent, nontoxic and environmentally safe herbicides. By using a knowledge-based approach, we identified physicochemical properties and substructures enriched in safe herbicides. By representing the small molecules as a graph, we leveraged these insights to guide the development of predictive models trained and tested on the largest collected data set of molecules with experimentally characterised herbicidal profiles to date (over 4500 compounds). In addition, we developed six new environmental and human toxicity predictors, spanning five different species to assist in molecule prioritisation. cropCSM was able to correctly identify 97% of herbicides currently available commercially, while predicting toxicity profiles with accuracies of up to 92%. We believe cropCSM will be an essential tool for the enrichment of screening libraries and to guide the development of potent and safe herbicides. We have made the method freely available through a user-friendly webserver at http://biosig.unimelb.edu.au/crop_csm.

Whole genome sequencing (WGS) can provide insight into drug-resistance, transmission chains and the identification of outbreaks, but data analysis remains an obstacle to its routine clinical use. Although several drug-resistance prediction tools have appeared, until now no website integrates drug-resistance prediction with strain genetic relationships and species identification of nontuberculous mycobacteria (NTM). We have established a free, function-rich, user-friendly online platform for MTB WGS data analysis (SAM-TB, http://samtb.szmbzx.com) that integrates drug-resistance prediction for 17 antituberculosis drugs, detection of variants, analysis of genetic relationships and NTM species identification. The accuracy of SAM-TB in predicting drug-resistance was assessed using 3177 sequenced clinical isolates with results of phenotypic drug-susceptibility tests (pDST). Compared to pDST, the sensitivity of SAM-TB for detecting multidrug-resistant tuberculosis was 93.9% [95% confidence interval (CI) 92.6-95.1%] with specificity of 96.2% (95% CI 95.2-97.1%). SAM-TB also analyzes the genetic relationships between multiple strains by reconstructing phylogenetic trees and calculating pairwise single nucleotide polymorphism (SNP) distances to identify genomic clusters. The incorporated mlstverse software identifies NTM species with an accuracy of 98.2% and Kraken2 software can detect mixed MTB and NTM samples. SAM-TB also has the capacity to share both sequence data and analysis between users. SAM-TB is a multifunctional integrated website that uses WGS raw data to accurately predict antituberculosis drug-resistance profiles, analyze genetic relationships between multiple strains and identify NTM species and mixed samples containing both NTM and MTB. SAM-TB is a useful tool for guiding both treatment and epidemiological investigation.

There is currently a transformed interest toward understanding the impact of fermentation on functional food development due to growing consumer interest on modified health benefits of sustainable foods. In this review, we attempt to summarize recent findings regarding the impact of Next-generation sequencing and other bioinformatics methods in the food microbiome and use prediction software to understand the critical role of microbes in producing fermented foods. Traditionally, fermentation methods and starter culture development were considered conventional methods needing optimization to eliminate errors in technique and were influenced by technical knowledge of fermentation. Recent advances in high-output omics innovations permit the implementation of additional logical tactics for developing fermentation methods. Further, the review describes the multiple functions of the predictions based on docking studies and the correlation of genomic and metabolomic analysis to develop trends to understand the potential food microbiome interactions and associated products to become a part of a healthy diet.

Unrestrained cellular growth and immune escape of a tumor are associated with the incidental errors of the genome and transcriptome. Advances in next-generation sequencing have identified thousands of genomic and transcriptomic aberrations that generate variant peptides that assemble the hidden proteome, further expanding the immunopeptidome. Emerging next-generation sequencing technologies and a number of computational methods estimated the abundance of immune infiltration from bulk transcriptome have advanced our understanding of tumor microenvironments. Here, we will characterize several major types of tumor-specific antigens arising from single-nucleotide variants, insertions and deletions, gene fusion, alternative splicing, RNA editing and non-coding RNAs. Finally, we summarize the current state-of-the-art computational and experimental approaches or resources and provide an integrative pipeline for the identification of candidate tumor antigens. Together, the systematic investigation of the hidden proteome in cancer will help facilitate the development of effective and durable immunotherapy targets for cancer.

Deep learning is a promising tool that uses nonlinear transformations to extract features from high-dimensional data. Deep learning is challenging in genome-wide association studies (GWAS) with high-dimensional genomic data. Here we propose a novel three-step approach (SWAT-CNN) for identification of genetic variants using deep learning to identify phenotype-related single nucleotide polymorphisms (SNPs) that can be applied to develop accurate disease classification models. In the first step, we divided the whole genome into nonoverlapping fragments of an optimal size and then ran convolutional neural network (CNN) on each fragment to select phenotype-associated fragments. In the second step, using a Sliding Window Association Test (SWAT), we ran CNN on the selected fragments to calculate phenotype influence scores (PIS) and identify phenotype-associated SNPs based on PIS. In the third step, we ran CNN on all identified SNPs to develop a classification model. We tested our approach using GWAS data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) including (N = 981; cognitively normal older adults (CN) = 650 and AD = 331). Our approach identified the well-known APOE region as the most significant genetic locus for AD. Our classification model achieved an area under the curve (AUC) of 0.82, which was compatible with traditional machine learning approaches, random forest and XGBoost. SWAT-CNN, a novel deep learning-based genome-wide approach, identified AD-associated SNPs and a classification model for AD and may hold promise for a range of biomedical applications.

Mass spectrometry-based proteomic technique has become indispensable in current exploration of complex and dynamic biological processes. Instrument development has largely ensured the effective production of proteomic data, which necessitates commensurate advances in statistical framework to discover the optimal proteomic signature. Current framework mainly emphasizes the generalizability of the identified signature in predicting the independent data but neglects the reproducibility among signatures identified from independently repeated trials on different sub-dataset. These problems seriously restricted the wide application of the proteomic technique in molecular biology and other related directions. Thus, it is crucial to enable the generalizable and reproducible discovery of the proteomic signature with the subsequent indication of phenotype association. However, no such tool has been developed and available yet. Herein, an online tool, POSREG, was therefore constructed to identify the optimal signature for a set of proteomic data. It works by (i) identifying the proteomic signature of good reproducibility and aggregating them to ensemble feature ranking by ensemble learning, (ii) assessing the generalizability of ensemble feature ranking to acquire the optimal signature and (iii) indicating the phenotype association of discovered signature. POSREG is unique in its capacity of discovering the proteomic signature by simultaneously optimizing its reproducibility and generalizability. It is now accessible free of charge without any registration or login requirement at https://idrblab.org/posreg/.

Although there are a large number of structural variations in the chromosomes of each individual, there is a lack of more accurate methods for identifying clinical pathogenic variants. Here, we proposed SVPath, a machine learning-based method to predict the pathogenicity of deletions, insertions and duplications structural variations that occur in exons. We constructed three types of annotation features for each structural variation event in the ClinVar database. First, we treated complex structural variations as multiple consecutive single nucleotide polymorphisms events, and annotated them with correlation scores based on single nucleic acid substitutions, such as the impact on protein function. Second, we determined which genes the variation occurred in, and constructed gene-based annotation features for each structural variation. Third, we also calculated related features based on the transcriptome, such as histone signal, the overlap ratio of variation and genomic element definitions, etc. Finally, we employed a gradient boosting decision tree machine learning method, and used the deletions, insertions and duplications in the ClinVar database to train a structural variation pathogenicity prediction model SVPath. These structural variations are clearly indicated as pathogenic or benign. Experimental results show that our SVPath has achieved excellent predictive performance and outperforms existing state-of-the-art tools. SVPath is very promising in evaluating the clinical pathogenicity of structural variants. SVPath can be used in clinical research to predict the clinical significance of unknown pathogenicity and new structural variation, so as to explore the relationship between diseases and structural variations in a computational way.

Protein secretion has a pivotal role in many biological processes and is particularly important for intercellular communication, from the cytoplasm to the host or external environment. Gram-positive bacteria can secrete proteins through multiple secretion pathways. The non-classical secretion pathway has recently received increasing attention among these secretion pathways, but its exact mechanism remains unclear. Non-classical secreted proteins (NCSPs) are a class of secreted proteins lacking signal peptides and motifs. Several NCSP predictors have been proposed to identify NCSPs and most of them employed the whole amino acid sequence of NCSPs to construct the model. However, the sequence length of different proteins varies greatly. In addition, not all regions of the protein are equally important and some local regions are not relevant to the secretion. The functional regions of the protein, particularly in the N- and C-terminal regions, contain important determinants for secretion. In this study, we propose a new hybrid deep learning-based framework, referred to as ASPIRER, which improves the prediction of NCSPs from amino acid sequences. More specifically, it combines a whole sequence-based XGBoost model and an N-terminal sequence-based convolutional neural network model; 5-fold cross-validation and independent tests demonstrate that ASPIRER achieves superior performance than existing state-of-the-art approaches. The source code and curated datasets of ASPIRER are publicly available at https://github.com/yanwu20/ASPIRER/. ASPIRER is anticipated to be a useful tool for improved prediction of novel putative NCSPs from sequences information and prioritization of candidate proteins for follow-up experimental validation.

The application of machine intelligence in biological sciences has led to the development of several automated tools, thus enabling rapid drug discovery. Adding to this development is the ongoing COVID-19 pandemic, due to which researchers working in the field of artificial intelligence have acquired an active interest in finding machine learning-guided solutions for diseases like mucormycosis, which has emerged as an important post-COVID-19 fungal complication, especially in immunocompromised patients. On these lines, we have proposed a temporal convolutional network-based binary classification approach to discover new antifungal molecules in the proteome of plants and animals to accelerate the development of antifungal medications. Although these biomolecules, known as antifungal peptides (AFPs), are part of an organism's intrinsic host defense mechanism, their identification and discovery by traditional biochemical procedures is arduous. Also, the absence of a large dataset on AFPs is also a considerable impediment in building a robust automated classifier. To this end, we have employed the transfer learning technique to pre-train our model on antibacterial peptides. Subsequently, we have built a classifier that predicts AFPs with accuracy and precision of 94%. Our classifier outperforms several state-of-the-art models by a considerable margin. The results of its performance were proven as statistically significant using the Kruskal-Wallis H test, followed by a post hoc analysis performed using the Tukey honestly significant difference (HSD) test. Furthermore, we identified potent AFPs in representative animal (Histatin) and plant (Snakin) proteins using our model. We also built and deployed a web app that is freely available at https://tcn-afppred.anvil.app/ for the identification of AFPs in protein sequences.

MOTIVATION: Identifying new therapeutic effects for the approved drugs is beneficial for effectively reducing the drug development cost and time. Most of the recent computational methods concentrate on exploiting multiple kinds of information about drugs and disease to predict the candidate associations between drugs and diseases. However, the drug and disease nodes have neighboring topologies with multiple scales, and the previous methods did not fully exploit and deeply integrate these topologies. RESULTS: We present a prediction method, multi-scale topology learning for drug-disease (MTRD), to integrate and learn multi-scale neighboring topologies and the attributes of a pair of drug and disease nodes. First, for multiple kinds of drug similarities, multiple drug-disease heterogenous networks are constructed respectively to integrate the similarities and associations related to drugs and diseases. Moreover, each heterogenous network has its specific topology structure, which is helpful for learning the corresponding specific topology representation. We formulate the topology embeddings for each drug node and disease node by random walking on each heterogeneous network, and the embeddings cover the neighboring topologies with different scopes. Because the multi-scale topology embeddings have context relationships, we construct Bi-directional long short-term memory-based module to encode these embeddings and their relationships and learn the neighboring topology representation. We also design the attention mechanisms at feature level and at scale level to obtain the more informative pairwise features and topology embeddings. A module based on multi-layer convolutional networks is constructed to learn the representative attributes of the drug-disease node pair according to their related similarity and association information. Comprehensive experimental results indicate that MTRD achieves the superior performance than several state-of-the-art methods for predicting drug-disease associations. MTRD also retrieves more actual drug-disease associations in the top-ranked candidates of the prediction result. Case studies on five drugs further demonstrate MTRD's ability in discovering the potential candidate diseases for the interested drugs.

MOTIVATION: Predicting disease-related long non-coding RNAs (lncRNAs) can be used as the biomarkers for disease diagnosis and treatment. The development of effective computational prediction approaches to predict lncRNA-disease associations (LDAs) can provide insights into the pathogenesis of complex human diseases and reduce experimental costs. However, few of the existing methods use microRNA (miRNA) information and consider the complex relationship between inter-graph and intra-graph in complex-graph for assisting prediction. RESULTS: In this paper, the relationships between the same types of nodes and different types of nodes in complex-graph are introduced. We propose a multi-channel graph attention autoencoder model to predict LDAs, called MGATE. First, an lncRNA-miRNA-disease complex-graph is established based on the similarity and correlation among lncRNA, miRNA and diseases to integrate the complex association among them. Secondly, in order to fully extract the comprehensive information of the nodes, we use graph autoencoder networks to learn multiple representations from complex-graph, inter-graph and intra-graph. Thirdly, a graph-level attention mechanism integration module is adopted to adaptively merge the three representations, and a combined training strategy is performed to optimize the whole model to ensure the complementary and consistency among the multi-graph embedding representations. Finally, multiple classifiers are explored, and Random Forest is used to predict the association score between lncRNA and disease. Experimental results on the public dataset show that the area under receiver operating characteristic curve and area under precision-recall curve of MGATE are 0.964 and 0.413, respectively. MGATE performance significantly outperformed seven state-of-the-art methods. Furthermore, the case studies of three cancers further demonstrate the ability of MGATE to identify potential disease-correlated candidate lncRNAs. The source code and supplementary data are available at https://github.com/sheng-n/MGATE. CONTACT: huanglan@jlu.edu.cn, wy6868@jlu.edu.cn.

Erratum in Brief Bioinform. 2022 Mar 15;:

The integration of multi-omics data makes it possible to understand complex biological organisms at the system level. Numerous integration approaches have been developed by assuming a common underlying data space. Due to the noise and heterogeneity of biological data, the performance of these approaches is greatly affected. In this work, we propose a novel deep neural network architecture, named Deep Latent Space Fusion (DLSF), which integrates the multi-omics data by learning consistent manifold in the sample latent space for disease subtypes identification. DLSF is built upon a cycle autoencoder with a shared self-expressive layer, which can naturally and adaptively merge nonlinear features at each omics level into one unified sample manifold and produce adaptive representation of heterogeneous samples at the multi-omics level. We have assessed DLSF on various biological and biomedical datasets to validate its effectiveness. DLSF can efficiently and accurately capture the intrinsic manifold of the sample structures or sample clusters compared with other state-of-the-art methods, and DLSF yielded more significant outcomes for biological significance, survival prognosis and clinical relevance in application of cancer study in The Cancer Genome Atlas. Notably, as a deep case study, we determined a new molecular subtype of kidney renal clear cell carcinoma that may benefit immunotherapy in the viewpoint of multi-omics, and we further found potential subtype-specific biomarkers from multiple omics data, which were validated by independent datasets. In addition, we applied DLSF to identify potential therapeutic agents of different molecular subtypes of chronic lymphocytic leukemia, demonstrating the scalability of DLSF in diverse omics data types and application scenarios.

Numerous experiments have demonstrated that abnormal expression of microRNAs (miRNAs) in organisms is often accompanied by the emergence of specific diseases. The research of miRNAs can promote the prevention and drug research of specific diseases. However, there are still many undiscovered links between miRNAs and diseases, which greatly limits the research of miRNAs. Therefore, for exploring the unknown miRNA-disease associations, we combine the graph random propagation network based on DropFeature with attention network to propose a novel deep learning model to predict the miRNA-disease associations (GRPAMDA). Specifically, we firstly construct the miRNA-disease heterogeneous graph based on miRNA-disease association information. Secondly, we adopt DropFeature to randomly delete the features of nodes in the graph and then perform propagation operations to enhance the features of miRNA and disease nodes. Thirdly, we employ the attention mechanism to fuse the features of random propagation by aggregating the enhanced neighbor features of miRNA and disease nodes. Finally, miRNA-disease association scores are generated by a fully connected layer. The average area under the curve of GRPAMDA model based on 5-fold cross-validation is 93.46% on HMDD v2.0. Case studies of esophageal tumors, lymphomas and prostate tumors show that 48, 47 and 46 of the top 50 miRNAs associated with these diseases are confirmed by dbDEMC and miR2Disease database, respectively. In short, the GRPAMDA model can be used as a valuable method to study miRNA-disease associations.

Multiple transcriptomic predictors of tumour cell radiosensitivity (RS) have been proposed, but they have not been benchmarked against one another or to control models. To address this, we present RadSigBench, a comprehensive benchmarking framework for RS signatures. The approach compares candidate models to those developed from randomly resampled control signatures and from cellular processes integral to the radiation response. Robust evaluation of signature accuracy, both overall and for individual tissues, is performed. The NCI60 and Cancer Cell Line Encyclopaedia datasets are integrated into our workflow. Prediction of two measures of RS is assessed: survival fraction after 2 Gy and mean inactivation dose. We apply the RadSigBench framework to seven prominent published signatures of radiation sensitivity and test for equivalence to control signatures. The mean out-of-sample R2 for the published models on test data was very poor at 0.01 (range: -0.05 to 0.09) for Cancer Cell Line Encyclopedia and 0.00 (range: -0.19 to 0.19) in the NCI60 data. The accuracy of both published and cellular process signatures investigated was equivalent to the resampled controls, suggesting that these signatures contain limited radiation-specific information. Enhanced modelling strategies are needed for effective prediction of intrinsic RS to inform clinical treatment regimes. We make recommendations for methodological improvements, for example the inclusion of perturbation data, multiomics, advanced machine learning and mechanistic modelling. Our validation framework provides for robust performance assessment of ongoing developments in intrinsic RS prediction.

Since the start of the SARS-CoV-2 pandemic in late 2019, several variants of concern (VOC) have been reported to have increased transmissibility. In addition, despite the progress of vaccination against SARS-CoV-2 worldwide, all vaccines currently in used are known to protect only partially from infection and onward transmission. We combined phylogenetic analysis with Bayesian inference under an epidemiological model to infer the reproduction number (Rt) and also trace person-to-person transmission. We examined the impact of phylogenetic uncertainty and sampling bias on the estimation. Our result indicated that lineage B had a significantly higher transmissibility than lineage A and contributed to the global pandemic to a large extent. In addition, although the transmissibility of VOCs is higher than other exponentially growing lineages, this difference is not very high. The probability of detecting onward transmission from patients infected with SARS-CoV-2 VOCs who had received at least one dose of vaccine was approximate 1.06% (3/284), which was slightly lower but not statistically significantly different from a probability of 1.21% (10/828) for unvaccinated individuals. In addition to VOCs, exponentially growing lineages in each country should also be account for when tailoring prevention and control strategies. One dose of vaccination could not efficiently prevent the onward transmission of SARS-CoV-2 VOCs. Consequently, nonpharmaceutical interventions (such as wearing masks and social distancing) should still be implemented in each country during the vaccination period.

Replication of DNA is an important process for the cell division cycle, gene expression regulation and other biological evolution processes. It also has a crucial role in a living organism's physical growth and structure. Replication of DNA comprises of three stages known as initiation, elongation and termination, whereas the origin of replication sites (ORI) is the location of initiation of the DNA replication process. There exist various methodologies to identify ORIs in the genomic sequences, however, these methods have used either extensive computations for execution, or have limited optimization for the large datasets. Herein, a model called ORI-Deep is proposed to identify ORIs from the multiple cell type genomic sequence benchmark data. An efficient method is proposed using a deep neural network to identify ORIs for four different eukaryotic species. For better representation of data, a feature vector is constructed using statistical moments for the training and testing of data and is further fed to a long short-term memory (LSTM) network. To prove the effectiveness of the proposed model, we applied several validation techniques at different levels to obtain seven accuracy metrics, and the accuracy score for self-consistency, 10-fold cross-validation, jackknife and the independent set test is observed to be 0.977, 0.948, 0.976 and 0.977, respectively. Based on the results, it can be concluded that ORI-Deep can efficiently predict the sites of origin replication in DNA sequence with high accuracy. Webserver for ORI-Deep is available at (https://share.streamlit.io/waqarhusain/orideep/main/app.py), whereas source code is available at (https://github.com/WaqarHusain/OriDeep).

Although drug combinations in cancer treatment appear to be a promising therapeutic strategy with respect to monotherapy, it is arduous to discover new synergistic drug combinations due to the combinatorial explosion. Deep learning technology holds immense promise for better prediction of in vitro synergistic drug combinations for certain cell lines. In methods applying such technology, omics data are widely adopted to construct cell line features. However, biological network data are rarely considered yet, which is worthy of in-depth study. In this study, we propose a novel deep learning method, termed PRODeepSyn, for predicting anticancer synergistic drug combinations. By leveraging the Graph Convolutional Network, PRODeepSyn integrates the protein-protein interaction (PPI) network with omics data to construct low-dimensional dense embeddings for cell lines. PRODeepSyn then builds a deep neural network with the Batch Normalization mechanism to predict synergy scores using the cell line embeddings and drug features. PRODeepSyn achieves the lowest root mean square error of 15.08 and the highest Pearson correlation coefficient of 0.75, outperforming two deep learning methods and four machine learning methods. On the classification task, PRODeepSyn achieves an area under the receiver operator characteristics curve of 0.90, an area under the precision-recall curve of 0.63 and a Cohen's Kappa of 0.53. In the ablation study, we find that using the multi-omics data and the integrated PPI network's information both can improve the prediction results. Additionally, the case study demonstrates the consistency between PRODeepSyn and previous studies.

Drug-target interactions (DTIs) prediction research presents important significance for promoting the development of modern medicine and pharmacology. Traditional biochemical experiments for DTIs prediction confront the challenges including long time period, high cost and high failure rate, and finally leading to a low-drug productivity. Chemogenomic-based computational methods can realize high-throughput prediction. In this study, we develop a deep collaborative filtering prediction model with multiembeddings, named DCFME (deep collaborative filtering prediction model with multiembeddings), which can jointly utilize multiple feature information from multiembeddings. Two different representation learning algorithms are first employed to extract heterogeneous network features. DCFME uses the generated low-dimensional dense vectors as input, and then simulates the drug-target relationship from the perspective of both couplings and heterogeneity. In addition, the model employs focal loss that concentrates the loss on sparse and hard samples in the training process. Comparative experiments with five baseline methods show that DCFME achieves more significant performance improvement on sparse datasets. Moreover, the model has better robustness and generalization capacity under several harder prediction scenarios.

Cancer is an omics disease. The development in high-throughput profiling has fundamentally changed cancer research and clinical practice. Compared with clinical, demographic and environmental data, the analysis of omics data-which has higher dimensionality, weaker signals and more complex distributional properties-is much more challenging. Developments in the literature are often 'scattered', with individual studies focused on one or a few closely related methods. The goal of this review is to assist cancer researchers with limited statistical expertise in establishing the 'overall framework' of cancer omics data analysis. To facilitate understanding, we mainly focus on intuition, concepts and key steps, and refer readers to the original publications for mathematical details. This review broadly covers unsupervised and supervised analysis, as well as individual-gene-based, gene-set-based and gene-network-based analysis. We also briefly discuss 'special topics' including interaction analysis, multi-datasets analysis and multi-omics analysis.

Protein-DNA interactions play crucial roles in the biological systems, and identifying protein-DNA binding sites is the first step for mechanistic understanding of various biological activities (such as transcription and repair) and designing novel drugs. How to accurately identify DNA-binding residues from only protein sequence remains a challenging task. Currently, most existing sequence-based methods only consider contextual features of the sequential neighbors, which are limited to capture spatial information. Based on the recent breakthrough in protein structure prediction by AlphaFold2, we propose an accurate predictor, GraphSite, for identifying DNA-binding residues based on the structural models predicted by AlphaFold2. Here, we convert the binding site prediction problem into a graph node classification task and employ a transformer-based variant model to take the protein structural information into account. By leveraging predicted protein structures and graph transformer, GraphSite substantially improves over the latest sequence-based and structure-based methods. The algorithm is further confirmed on the independent test set of 181 proteins, where GraphSite surpasses the state-of-the-art structure-based method by 16.4% in area under the precision-recall curve and 11.2% in Matthews correlation coefficient, respectively. We provide the datasets, the predicted structures and the source codes along with the pre-trained models of GraphSite at https://github.com/biomed-AI/GraphSite. The GraphSite web server is freely available at https://biomed.nscc-gz.cn/apps/GraphSite.

Predicting drug-target interactions (DTIs) is a convenient strategy for drug discovery. Although various computational methods have been put forward in recent years, DTIs prediction is still a challenging task. In this paper, based on indirect prior information (we term them as mediators), we proposed a new model, called Bridging-BPs (bridging paths), for DTIs prediction. Specifically, we regarded linkage process between mediators and DTs (drugs and proteins) as 'bridging' and source (drug)-mediators-destination (protein) as bridging paths. By integrating various bridging paths, we constructed a bridging heterogeneous graph for DTIs. After that, an improved graph-embedding algorithm-BPs2vec-was designed to capture deep topological features underlying the bridging graph, thereby obtaining the low-dimensional node vector representations. Then, the vector representations were fed into a Random Forest classifier to train and score the probability, outputting the final classification results for potential DTIs. Under 5-fold cross validation, our method obtained AUPR of 88.97% and AUC of 88.63%, suggesting that Bridging-BPs could effectively mine the link relationships hidden in indirect prior information and it significantly improved the accuracy and robustness of DTIs prediction without direct prior information. Finally, we confirmed the practical prediction ability of Bridging-BPs by case studies.

As an important post-translational modification, lysine ubiquitination participates in numerous biological processes and is involved in human diseases, whereas the site specificity of ubiquitination is mainly decided by ubiquitin-protein ligases (E3s). Although numerous ubiquitination predictors have been developed, computational prediction of E3-specific ubiquitination sites is still a great challenge. Here, we carefully reviewed the existing tools for the prediction of general ubiquitination sites. Also, we developed a tool named GPS-Uber for the prediction of general and E3-specific ubiquitination sites. From the literature, we manually collected 1311 experimentally identified site-specific E3-substrate relations, which were classified into different clusters based on corresponding E3s at different levels. To predict general ubiquitination sites, we integrated 10 types of sequence and structure features, as well as three types of algorithms including penalized logistic regression, deep neural network and convolutional neural network. Compared with other existing tools, the general model in GPS-Uber exhibited a highly competitive accuracy, with an area under curve values of 0.7649. Then, transfer learning was adopted for each E3 cluster to construct E3-specific models, and in total 112 individual E3-specific predictors were implemented. Using GPS-Uber, we conducted a systematic prediction of human cancer-associated ubiquitination events, which could be helpful for further experimental consideration. GPS-Uber will be regularly updated, and its online service is free for academic research at http://gpsuber.biocuckoo.cn/.

Two-dimensional gas chromatography-time-of-flight mass spectrometry (GC × GC-TOFMS) provides a large amount of molecular information from biological samples. However, the lack of a comprehensive compound library or customizable bioinformatics tool is currently a challenge in GC × GC-TOFMS data analysis. We present an open-source deep learning (DL) software called contour regions of interest (ROI) identification, simulation and untargeted metabolomics profiler (CRISP). CRISP integrates multiple customizable deep neural network architectures for assisting the semi-automated identification of ROIs, contour synthesis, resolution enhancement and classification of GC × GC-TOFMS-based contour images. The approach includes the novel aggregate feature representative contour (AFRC) construction and stacked ROIs. This generates an unbiased contour image dataset that enhances the contrasting characteristics between different test groups and can be suitable for small sample sizes. The utility of the generative models and the accuracy and efficacy of the platform were demonstrated using a dataset of GC × GC-TOFMS contour images from patients with late-stage diabetic nephropathy and healthy control groups. CRISP successfully constructed AFRC images and identified over five ROIs to create a deepstacked dataset. The high fidelity, 512 × 512-pixels generative model was trained as a generator with a Fréchet inception distance of <47.00. The trained classifier achieved an AUROC of >0.96 and a classification accuracy of >95.00% for datasets with and without column bleed. Overall, CRISP demonstrates good potential as a DL-based approach for the rapid analysis of 4-D GC × GC-TOFMS untargeted metabolite profiles by directly implementing contour images. CRISP is available at https://github.com/vivekmathema/GCxGC-CRISP.

Promoters are crucial regulatory DNA regions for gene transcriptional activation. Rapid advances in next-generation sequencing technologies have accelerated the accumulation of genome sequences, providing increased training data to inform computational approaches for both prokaryotic and eukaryotic promoter prediction. However, it remains a significant challenge to accurately identify species-specific promoter sequences using computational approaches. To advance computational support for promoter prediction, in this study, we curated 58 comprehensive, up-to-date, benchmark datasets for 7 different species (i.e. Escherichia coli, Bacillus subtilis, Homo sapiens, Mus musculus, Arabidopsis thaliana, Zea mays and Drosophila melanogaster) to assist the research community to assess the relative functionality of alternative approaches and support future research on both prokaryotic and eukaryotic promoters. We revisited 106 predictors published since 2000 for promoter identification (40 for prokaryotic promoter, 61 for eukaryotic promoter, and 5 for both). We systematically evaluated their training datasets, computational methodologies, calculated features, performance and software usability. On the basis of these benchmark datasets, we benchmarked 19 predictors with functioning webservers/local tools and assessed their prediction performance. We found that deep learning and traditional machine learning-based approaches generally outperformed scoring function-based approaches. Taken together, the curated benchmark dataset repository and the benchmarking analysis in this study serve to inform the design and implementation of computational approaches for promoter prediction and facilitate more rigorous comparison of new techniques in the future.

Predicting the difference in thermodynamic stability between protein variants is crucial for protein design and understanding the genotype-phenotype relationships. So far, several computational tools have been created to address this task. Nevertheless, most of them have been trained or optimized on the same and 'all' available data, making a fair comparison unfeasible. Here, we introduce a novel dataset, collected and manually cleaned from the latest version of the ThermoMutDB database, consisting of 669 variants not included in the most widely used training datasets. The prediction performance and the ability to satisfy the antisymmetry property by considering both direct and reverse variants were evaluated across 21 different tools. The Pearson correlations of the tested tools were in the ranges of 0.21-0.5 and 0-0.45 for the direct and reverse variants, respectively. When both direct and reverse variants are considered, the antisymmetric methods perform better achieving a Pearson correlation in the range of 0.51-0.62. The tested methods seem relatively insensitive to the physiological conditions, performing well also on the variants measured with more extreme pH and temperature values. A common issue with all the tested methods is the compression of the $\Delta \Delta G$ predictions toward zero. Furthermore, the thermodynamic stability of the most significantly stabilizing variants was found to be more challenging to predict. This study is the most extensive comparisons of prediction methods using an entirely novel set of variants never tested before.

Spatial structures of proteins are closely related to protein functions. Integrating protein structures improves the performance of protein-protein interaction (PPI) prediction. However, the limited quantity of known protein structures restricts the application of structure-based prediction methods. Utilizing the predicted protein structure information is a promising method to improve the performance of sequence-based prediction methods. We propose a novel end-to-end framework, TAGPPI, to predict PPIs using protein sequence alone. TAGPPI extracts multi-dimensional features by employing 1D convolution operation on protein sequences and graph learning method on contact maps constructed from AlphaFold. A contact map contains abundant spatial structure information, which is difficult to obtain from 1D sequence data directly. We further demonstrate that the spatial information learned from contact maps improves the ability of TAGPPI in PPI prediction tasks. We compare the performance of TAGPPI with those of nine state-of-the-art sequence-based methods, and TAGPPI outperforms such methods in all metrics. To the best of our knowledge, this is the first method to use the predicted protein topology structure graph for sequence-based PPI prediction. More importantly, our proposed architecture could be extended to other prediction tasks related to proteins.

Since miRNAs can participate in the posttranscriptional regulation of gene expression, they may provide ideas for the development of new drugs or become new biomarkers for drug targets or disease diagnosis. In this work, we propose an miRNA-disease association prediction method based on meta-paths (MDPBMP). First, an miRNA-disease-gene heterogeneous information network was constructed, and seven symmetrical meta-paths were defined according to different semantics. After constructing the initial feature vector for the node, the vector information carried by all nodes on the meta-path instance is extracted and aggregated to update the feature vector of the starting node. Then, the vector information obtained by the nodes on different meta-paths is aggregated. Finally, miRNA and disease embedding feature vectors are used to calculate their associated scores. Compared with the other methods, MDPBMP obtained the highest AUC value of 0.9214. Among the top 50 predicted miRNAs for lung neoplasms, esophageal neoplasms, colon neoplasms and breast neoplasms, 49, 48, 49 and 50 have been verified. Furthermore, for breast neoplasms, we deleted all the known associations between breast neoplasms and miRNAs from the training set. These results also show that for new diseases without known related miRNA information, our model can predict their potential miRNAs. Code and data are available at https://github.com/LiangYu-Xidian/MDPBMP.

Advancement in single-cell RNA sequencing leads to exponential accumulation of single-cell expression data. However, there is still lack of tools that could integrate these unlimited accumulations of single-cell expression data. Here, we presented a universal approach iSEEEK for integrating super large-scale single-cell expression via exploring expression rankings of top-expressing genes. We developed iSEEEK with 11.9 million single cells. We demonstrated the efficiency of iSEEEK with canonical single-cell downstream tasks on five heterogenous datasets encompassing human and mouse samples. iSEEEK achieved good clustering performance benchmarked against well-annotated cell labels. In addition, iSEEEK could transfer its knowledge learned from large-scale expression data on new dataset that was not involved in its development. iSEEEK enables identification of gene-gene interaction networks that are characteristic of specific cell types. Our study presents a simple and yet effective method to integrate super large-scale single-cell transcriptomes and would facilitate translational single-cell research from bench to bedside.