Gene regulatory networks (GRNs) reveal the complex molecular interactions that govern cell state. However, it is challenging for identifying causal relations among genes due to noisy data and molecular nonlinearity. Here, we propose a novel causal criterion, neighbor cross-mapping entropy (NME), for inferring GRNs from both steady data and time-series data. NME is designed to quantify 'continuous causality' or functional dependency from one variable to another based on their function continuity with varying neighbor sizes. NME shows superior performance on benchmark datasets, comparing with existing methods. By applying to scRNA-seq datasets, NME not only reliably inferred GRNs for cell types but also identified cell states. Based on the inferred GRNs and further their activity matrices, NME showed better performance in single-cell clustering and downstream analyses. In summary, based on continuous causality, NME provides a powerful tool in inferring causal regulations of GRNs between genes from scRNA-seq data, which is further exploited to identify novel cell types/states and predict cell type-specific network modules.

With the development of chromosome conformation capture technology, the genome-wide investigation of higher-order chromatin structure by using high-throughput chromatin conformation capture (Hi-C) technology is emerging as an important component for understanding the mechanism of gene regulation. Considering genetic and epigenetic differences are typically used to explore the pathological reasons on the chromosome and gene level, visualizing multi-omics data and performing an intuitive analysis by using an interactive browser become a powerful and welcomed way. In this paper, we develop an effective sequence and chromatin interaction data display browser called HiBrowser for visualizing and analyzing Hi-C data and their associated genetic and epigenetic annotations. The advantages of HiBrowser are flexible multi-omics navigation, novel multidimensional synchronization comparisons and dynamic interaction system. In particular, HiBrowser first provides an out of the box web service and allows flexible and dynamic reconstruction of custom annotation tracks on demand during running. In order to conveniently and intuitively analyze the similarities and differences among multiple samples, such as visual comparisons of normal and tumor tissue samples, and pan genomes of multiple (consanguineous) species, HiBrowser develops a clone mode to synchronously display the genome coordinate positions or the same regions of multiple samples on the same page of visualization. HiBrowser also supports a pluralistic and precise search on correlation data of distal cis-regulatory elements and navigation to any region on Hi-C heatmap of interest according to the searching results. HiBrowser is a no-build tool, and could be easily deployed in local server. The source code is available at https://github.com/lyotvincent/HiBrowser.

Prediction of driver genes (tumor suppressors and oncogenes) is an essential step in understanding cancer development and discovering potential novel treatments. We recently proposed Moonlight as a bioinformatics framework to predict driver genes and analyze them in a system-biology-oriented manner based on -omics integration. Moonlight uses gene expression as a primary data source and combines it with patterns related to cancer hallmarks and regulatory networks to identify oncogenic mediators. Once the oncogenic mediators are identified, it is important to include extra levels of evidence, called mechanistic indicators, to identify driver genes and to link the observed gene expression changes to the underlying alteration that promotes them. Such a mechanistic indicator could be for example a mutation in the regulatory regions for the candidate gene. Here, we developed new functionalities and released Moonlight2 to provide the user with a mutation-based mechanistic indicator as a second layer of evidence. These functionalities analyze mutations in a cancer cohort to classify them into driver and passenger mutations. Those oncogenic mediators with at least one driver mutation are retained as the final set of driver genes. We applied Moonlight2 to the basal-like breast cancer subtype, lung adenocarcinoma and thyroid carcinoma using data from The Cancer Genome Atlas. For example, in basal-like breast cancer, we found four oncogenes (COPZ2, SF3B4, KRTCAP2 and POLR2J) and nine tumor suppressor genes (KIR2DL4, KIF26B, ARL15, ARHGAP25, EMCN, GMFG, TPK1, NR5A2 and TEK) containing a driver mutation in their promoter region, possibly explaining their deregulation. Moonlight2R is available at https://github.com/ELELAB/Moonlight2R.

MOTIVATION: Recent advances in spatially resolved transcriptomics (ST) technologies enable the measurement of gene expression profiles while preserving cellular spatial context. Linking gene expression of cells with their spatial distribution is essential for better understanding of tissue microenvironment and biological progress. However, effectively combining gene expression data with spatial information to identify spatial domains remains challenging. RESULTS: To deal with the above issue, in this paper, we propose a novel unsupervised learning framework named STMGCN for identifying spatial domains using multi-view graph convolution networks (MGCNs). Specifically, to fully exploit spatial information, we first construct multiple neighbor graphs (views) with different similarity measures based on the spatial coordinates. Then, STMGCN learns multiple view-specific embeddings by combining gene expressions with each neighbor graph through graph convolution networks. Finally, to capture the importance of different graphs, we further introduce an attention mechanism to adaptively fuse view-specific embeddings and thus derive the final spot embedding. STMGCN allows for the effective utilization of spatial context to enhance the expressive power of the latent embeddings with multiple graph convolutions. We apply STMGCN on two simulation datasets and five real spatial transcriptomics datasets with different resolutions across distinct platforms. The experimental results demonstrate that STMGCN obtains competitive results in spatial domain identification compared with five state-of-the-art methods, including spatial and non-spatial alternatives. Besides, STMGCN can detect spatially variable genes with enriched expression patterns in the identified domains. Overall, STMGCN is a powerful and efficient computational framework for identifying spatial domains in spatial transcriptomics data.

Combination therapies have brought significant advancements to the treatment of various diseases in the medical field. However, searching for effective drug combinations remains a major challenge due to the vast number of possible combinations. Biomedical knowledge graph (KG)-based methods have shown potential in predicting effective combinations for wide spectrum of diseases, but the lack of credible negative samples has limited the prediction performance of machine learning models. To address this issue, we propose a novel model-agnostic framework that leverages existing drug-drug interaction (DDI) data as a reliable negative dataset and employs supervised contrastive learning (SCL) to transform drug embedding vectors to be more suitable for drug combination prediction. We conducted extensive experiments using various network embedding algorithms, including random walk and graph neural networks, on a biomedical KG. Our framework significantly improved performance metrics compared to the baseline framework. We also provide embedding space visualizations and case studies that demonstrate the effectiveness of our approach. This work highlights the potential of using DDI data and SCL in finding tighter decision boundaries for predicting effective drug combinations.

Integrating single-cell multi-omics data is a challenging task that has led to new insights into complex cellular systems. Various computational methods have been proposed to effectively integrate these rapidly accumulating datasets, including deep learning. However, despite the proven success of deep learning in integrating multi-omics data and its better performance over classical computational methods, there has been no systematic study of its application to single-cell multi-omics data integration. To fill this gap, we conducted a literature review to explore the use of multimodal deep learning techniques in single-cell multi-omics data integration, taking into account recent studies from multiple perspectives. Specifically, we first summarized different modalities found in single-cell multi-omics data. We then reviewed current deep learning techniques for processing multimodal data and categorized deep learning-based integration methods for single-cell multi-omics data according to data modality, deep learning architecture, fusion strategy, key tasks and downstream analysis. Finally, we provided insights into using these deep learning models to integrate multi-omics data and better understand single-cell biological mechanisms.

MicroRNAs (miRNAs) silence genes by binding to messenger RNAs, whereas long non-coding RNAs (lncRNAs) act as competitive endogenous RNAs (ceRNAs) that can relieve miRNA silencing effects and upregulate target gene expression. The ceRNA association between lncRNAs and miRNAs has been a research hotspot due to its medical importance, but it is challenging to verify experimentally. In this paper, we propose a novel deep learning scheme, i.e. sequence pre-training-based graph neural network (SPGNN), that combines pre-training and fine-tuning stages to predict lncRNA-miRNA associations from RNA sequences and the existing interactions represented as a graph. First, we utilize a sequence-to-vector technique to generate pre-trained embeddings based on the sequences of all RNAs during the pre-training stage. In the fine-tuning stage, we use Graph Neural Network to learn node representations from the heterogeneous graph constructed using lncRNA-miRNA association information. We evaluate our proposed scheme SPGNN on our newly collected animal lncRNA-miRNA association dataset and demonstrate that combining the $k$-mer technique and Doc2vec model for pre-training with the Simple Graph Convolution Network for fine-tuning is effective in predicting lncRNA-miRNA associations. Our approach outperforms state-of-the-art baselines across various evaluation metrics. We also conduct an ablation study and hyperparameter analysis to verify the effectiveness of each component and parameter of our scheme. The complete code and dataset are available on GitHub: https://github.com/zixwang/SPGNN.

The accurate prediction of the effect of amino acid mutations for protein-protein interactions (PPI $\Delta \Delta G$) is a crucial task in protein engineering, as it provides insight into the relevant biological processes underpinning protein binding and provides a basis for further drug discovery. In this study, we propose MpbPPI, a novel multi-task pre-training-based geometric equivariance-preserving framework to predict PPI  $\Delta \Delta G$. Pre-training on a strictly screened pre-training dataset is employed to address the scarcity of protein-protein complex structures annotated with PPI $\Delta \Delta G$ values. MpbPPI employs a multi-task pre-training technique, forcing the framework to learn comprehensive backbone and side chain geometric regulations of protein-protein complexes at different scales. After pre-training, MpbPPI can generate high-quality representations capturing the effective geometric characteristics of labeled protein-protein complexes for downstream $\Delta \Delta G$ predictions. MpbPPI serves as a scalable framework supporting different sources of mutant-type (MT) protein-protein complexes for flexible application. Experimental results on four benchmark datasets demonstrate that MpbPPI is a state-of-the-art framework for PPI $\Delta \Delta G$ predictions. The data and source code are available at https://github.com/arantir123/MpbPPI.

Single-cell high-throughput chromatin conformation capture technologies (scHi-C) has been used to map chromatin spatial organization in complex tissues. However, computational tools to detect differential chromatin contacts (DCCs) from scHi-C datasets in development and through disease pathogenesis are still lacking. Here, we present SnapHiC-D, a computational pipeline to identify DCCs between two scHi-C datasets. Compared to methods designed for bulk Hi-C data, SnapHiC-D detects DCCs with high sensitivity and accuracy. We used SnapHiC-D to identify cell-type-specific chromatin contacts at 10 Kb resolution in mouse hippocampal and human prefrontal cortical tissues, demonstrating that DCCs detected in the hippocampal and cortical cell types are generally associated with cell-type-specific gene expression patterns and epigenomic features. SnapHiC-D is freely available at https://github.com/HuMingLab/SnapHiC-D.

Protein function prediction based on amino acid sequence alone is an extremely challenging but important task, especially in metagenomics/metatranscriptomics field, in which novel proteins have been uncovered exponentially from new microorganisms. Many of them are extremely low homology to known proteins and cannot be annotated with homology-based or information integrative methods. To overcome this problem, we proposed a Homology Independent protein Function annotation method (HiFun) based on a unified deep-learning model by reassembling the sequence as protein language. The robustness of HiFun was evaluated using the benchmark datasets and metrics in the CAFA3 challenge. To navigate the utility of HiFun, we annotated 2 212 663 unknown proteins and discovered novel motifs in the UHGP-50 catalog. We proved that HiFun can extract latent function related structure features which empowers it ability to achieve function annotation for non-homology proteins. HiFun can substantially improve newly proteins annotation and expand our understanding of microorganisms' adaptation in various ecological niches. Moreover, we provided a free and accessible webservice at http://www.unimd.org/HiFun, requiring only protein sequences as input, offering researchers an efficient and practical platform for predicting protein functions.

Protein crystallization is crucial for biology, but the steps involved are complex and demanding in terms of external factors and internal structure. To save on experimental costs and time, the tendency of proteins to crystallize can be initially determined and screened by modeling. As a result, this study created a new pipeline aimed at using protein sequence to predict protein crystallization propensity in the protein material production stage, purification stage and production of crystal stage. The newly created pipeline proposed a new feature selection method, which involves combining Chi-square (${\chi }^{2}$) and recursive feature elimination together with the 12 selected features, followed by a linear discriminant analysisfor dimensionality reduction and finally, a support vector machine algorithm with hyperparameter tuning and 10-fold cross-validation is used to train the model and test the results. This new pipeline has been tested on three different datasets, and the accuracy rates are higher than the existing pipelines. In conclusion, our model provides a new solution to predict multistage protein crystallization propensity which is a big challenge in computational biology.

Cancer driver genes are critical in driving tumor cell growth, and precisely identifying these genes is crucial in advancing our understanding of cancer pathogenesis and developing targeted cancer drugs. Despite the current methods for discovering cancer driver genes that mainly rely on integrating multi-omics data, many existing models are overly complex, and it is difficult to interpret the results accurately. This study aims to address this issue by introducing InDEP, an interpretable machine learning framework based on cascade forests. InDEP is designed with easy-to-interpret features, cascade forests based on decision trees and a KernelSHAP module that enables fine-grained post-hoc interpretation. Integrating multi-omics data, InDEP can identify essential features of classified driver genes at both the gene and cancer-type levels. The framework accurately identifies driver genes, discovers new patterns that make genes as driver genes and refines the cancer driver gene catalog. In comparison with state-of-the-art methods, InDEP proved to be more accurate on the test set and identified reliable candidate driver genes. Mutational features were the primary drivers for InDEP's identifying driver genes, with other omics features also contributing. At the gene level, the framework concluded that substitution-type mutations were the main reason most genes were identified as driver genes. InDEP's ability to identify reliable candidate driver genes opens up new avenues for precision oncology and discovering new biomedical knowledge. This framework can help advance cancer research by providing an interpretable method for identifying cancer driver genes and their contribution to cancer pathogenesis, facilitating the development of targeted cancer drugs.

Cancer heterogeneity has posed great challenges in exploring precise therapeutic strategies for cancer treatment. The identification of cancer subtypes aims to detect patients with distinct molecular profiles and thus could provide new clues on effective clinical therapies. While great efforts have been made, it remains challenging to develop powerful computational methods that can efficiently integrate multi-omics datasets for the task. In this paper, we propose a novel self-supervised learning model called Deep Multi-view Contrastive Learning (DMCL) for cancer subtype identification. Specifically, by incorporating the reconstruction loss, contrastive loss and clustering loss into a unified framework, our model simultaneously encodes the sample discriminative information into the extracted feature representations and well preserves the sample cluster structures in the embedded space. Moreover, DMCL is an end-to-end framework where the cancer subtypes could be directly obtained from the model outputs. We compare DMCL with eight alternatives ranging from classic cancer subtype identification methods to recently developed state-of-the-art systems on 10 widely used cancer multi-omics datasets as well as an integrated dataset, and the experimental results validate the superior performance of our method. We further conduct a case study on liver cancer and the analysis results indicate that different subtypes might have different responses to the selected chemotherapeutic drugs.

Duplex sequencing technology has been widely used in the detection of low-frequency mutations in circulating tumor deoxyribonucleic acid (DNA), but how to determine the sequencing depth and other experimental parameters to ensure the stable detection of low-frequency mutations is still an urgent problem to be solved. The mutation detection rules of duplex sequencing constrain not only the number of mutated templates but also the number of mutation-supportive reads corresponding to each forward and reverse strand of the mutated templates. To tackle this problem, we proposed a Depth Estimation model for stable detection of Low-Frequency MUTations in duplex sequencing (DELFMUT), which models the identity correspondence and quantitative relationships between templates and reads using the zero-truncated negative binomial distribution without considering the sequences composed of bases. The results of DELFMUT were verified by real duplex sequencing data. In the case of known mutation frequency and mutation detection rule, DELFMUT can recommend the combinations of DNA input and sequencing depth to guarantee the stable detection of mutations, and it has a great application value in guiding the experimental parameter setting of duplex sequencing technology.

Enhancers are crucial cis-regulatory elements that control gene expression in a cell-type-specific manner. Despite extensive genetic and computational studies, accurately predicting enhancer activity in different cell types remains a challenge, and the grammar of enhancers is still poorly understood. Here, we present HEAP (high-resolution enhancer activity prediction), an explainable deep learning framework for predicting enhancers and exploring enhancer grammar. The framework includes three modules that use grammar-based reasoning for enhancer prediction. The algorithm can incorporate DNA sequences and epigenetic modifications to obtain better accuracy. We use a novel two-step multi-task learning method, task adaptive parameter sharing (TAPS), to efficiently predict enhancers in different cell types. We first train a shared model with all cell-type datasets. Then we adapt to specific tasks by adding several task-specific subset layers. Experiments demonstrate that HEAP outperforms published methods and showcases the effectiveness of the TAPS, especially for those with limited training samples. Notably, the explainable framework HEAP utilizes post-hoc interpretation to provide insights into the prediction mechanisms from three perspectives: data, model architecture and algorithm, leading to a better understanding of model decisions and enhancer grammar. To the best of our knowledge, HEAP will be a valuable tool for insight into the complex mechanisms of enhancer activity.

The three-dimensional organization of chromatin plays a crucial role in gene regulation and cellular processes like deoxyribonucleic acid (DNA) transcription, replication and repair. Hi-C and related techniques provide detailed views of spatial proximities within the nucleus. However, data analysis is challenging partially due to a lack of well-defined, underpinning mathematical frameworks. Recently, recognizing and analyzing geometric patterns in Hi-C data has emerged as a powerful approach. This review provides a summary of algorithms for automatic recognition and analysis of geometric patterns in Hi-C data and their correspondence with chromatin structure. We classify existing algorithms on the basis of the data representation and pattern recognition paradigm they make use of. Finally, we outline some of the challenges ahead and promising future directions.

Silencers are noncoding DNA sequence fragments located on the genome that suppress gene expression. The variation of silencers in specific cells is closely related to gene expression and cancer development. Computational approaches that exclusively rely on DNA sequence information for silencer identification fail to account for the cell specificity of silencers, resulting in diminished accuracy. Despite the discovery of several transcription factors and epigenetic modifications associated with silencers on the genome, there is still no definitive biological signal or combination thereof to fully characterize silencers, posing challenges in selecting suitable biological signals for their identification. Therefore, we propose a sophisticated deep learning framework called DeepICSH, which is based on multiple biological data sources. Specifically, DeepICSH leverages a deep convolutional neural network to automatically capture biologically relevant signal combinations strongly associated with silencers, originating from a diverse array of biological signals. Furthermore, the utilization of attention mechanisms facilitates the scoring and visualization of these signal combinations, whereas the employment of skip connections facilitates the fusion of multilevel sequence features and signal combinations, thereby empowering the accurate identification of silencers within specific cells. Extensive experiments on HepG2 and K562 cell line data sets demonstrate that DeepICSH outperforms state-of-the-art methods in silencer identification. Notably, we introduce for the first time a deep learning framework based on multi-omics data for classifying strong and weak silencers, achieving favorable performance. In conclusion, DeepICSH shows great promise for advancing the study and analysis of silencers in complex diseases. The source code is available at https://github.com/lyli1013/DeepICSH.

RNA-binding proteins (RBPs) are central actors of RNA post-transcriptional regulation. Experiments to profile-binding sites of RBPs in vivo are limited to transcripts expressed in the experimental cell type, creating the need for computational methods to infer missing binding information. While numerous machine-learning based methods have been developed for this task, their use of heterogeneous training and evaluation datasets across different sets of RBPs and CLIP-seq protocols makes a direct comparison of their performance difficult. Here, we compile a set of 37 machine learning (primarily deep learning) methods for in vivo RBP-RNA interaction prediction and systematically benchmark a subset of 11 representative methods across hundreds of CLIP-seq datasets and RBPs. Using homogenized sample pre-processing and two negative-class sample generation strategies, we evaluate methods in terms of predictive performance and assess the impact of neural network architectures and input modalities on model performance. We believe that this study will not only enable researchers to choose the optimal prediction method for their tasks at hand, but also aid method developers in developing novel, high-performing methods by introducing a standardized framework for their evaluation.

Microhomology-mediated end joining (MMEJ), an error-prone DNA damage repair mechanism, frequently leads to chromosomal rearrangements due to its ability to engage in promiscuous end joining of genomic instability and also leads to increasing mutational load at the sequences flanking the breakpoints (BPs). In this study, we systematically investigated the homology sequences around the genomic breakpoint area of human fusion genes, which were formed by the chromosomal rearrangements initiated by DNA double-strand breakage. Since the RNA-seq data is the typical data set to check the fusion genes, for the known exon junction fusion breakpoints identified from RNA-seq data, we have to infer the high chance of genomic breakpoint regions. For this, we utilized the high feature importance score area calculated from our recently developed fusion BP prediction model, FusionAI and identified 151 K microhomologies among ~24 K fusion BPs in 20 K fusion genes. From our multiple bioinformatics studies, we found a relationship between sequence homologies and the immune system. This in-silico study will provide novel knowledge on the sequence homologies around the coded structural variants.

Non-coding RNA (ncRNA) plays a critical role in biology. ncRNAs from the same family usually have similar functions, as a result, it is essential to predict ncRNA families before identifying their functions. There are two primary methods for predicting ncRNA families, namely, traditional biological methods and computational methods. In traditional biological methods, a lot of manpower and resources are required to predict ncRNA families. Therefore, this paper proposed a new ncRNA family prediction method called MFPred based on computational methods. MFPred identified ncRNA families by extracting sequence features of ncRNAs, and it possessed three primary modules, including (1) four ncRNA sequences encoding and feature extraction module, which encoded ncRNA sequences and extracted four different features of ncRNA sequences, (2) dynamic Bi_GRU and feature fusion module, which extracted contextual information features of the ncRNA sequence and (3) ResNet_SE module that extracted local information features of the ncRNA sequence. In this study, MFPred was compared with the previously proposed ncRNA family prediction methods using two frequently used public ncRNA datasets, NCY and nRC. The results showed that MFPred outperformed other prediction methods in the two datasets.

Ion mobility coupled to mass spectrometry informs on the shape and size of protein structures in the form of a collision cross section (CCSIM). Although there are several computational methods for predicting CCSIM based on protein structures, including our previously developed projection approximation using rough circular shapes (PARCS), the process usually requires prior experience with the command-line interface. To overcome this challenge, here we present a web application on the Rosetta Online Server that Includes Everyone (ROSIE) webserver to predict CCSIM from protein structure using projection approximation with PARCS. In this web interface, the user is only required to provide one or more PDB files as input. Results from our case studies suggest that CCSIM predictions (with ROSIE-PARCS) are highly accurate with an average error of 6.12%. Furthermore, the absolute difference between CCSIM and CCSPARCS can help in distinguishing accurate from inaccurate AlphaFold2 protein structure predictions. ROSIE-PARCS is designed with a user-friendly interface, is available publicly and is free to use. The ROSIE-PARCS web interface is supported by all major web browsers and can be accessed via this link (https://rosie.graylab.jhu.edu).

The formation of biomolecular condensates by liquid-liquid phase separation (LLPS) has become a universal mechanism for spatiotemporal coordination of biological activities in cells and has been widely observed to directly regulate the key cellular processes involved in cancer cell pathology. However, the complexity of protein sequences and the diversity of conformations are inherently disordered, which poses great challenges for LLPS protein calculations and experimental research. Herein, we proposed a novel predictor named PredLLPS_PSSM for LLPS protein identification based only on sequence evolution information. Because finding real and reliable samples is the cornerstone of building predictors, we collected anew and collated the LLPS proteins from the latest versions of three databases. By comparing the performance of the position-specific score matrix (PSSM) and word embedding, PredLLPS_PSSM combined PSSM-based information and two deep learning frameworks. Independent tests using three existing independent test datasets and two newly constructed independent test datasets demonstrated the superiority of PredLLPS_PSSM compared with state-of-the-art methods. Furthermore, we tested PredLLPS_PSSM on nine experimentally identified LLPS proteins from three insects that were not included in any of the databases. In addition, the powerful Shapley Additive exPlanation algorithm and heatmap were applied to find the most critical amino acids relevant to LLPS.

Predicting the biological properties of molecules is crucial in computer-aided drug development, yet it's often impeded by data scarcity and imbalance in many practical applications. Existing approaches are based on self-supervised learning or 3D data and using an increasing number of parameters to improve performance. These approaches may not take full advantage of established chemical knowledge and could inadvertently introduce noise into the respective model. In this study, we introduce a more elegant transformer-based framework with focused attention for molecular representation (TransFoxMol) to improve the understanding of artificial intelligence (AI) of molecular structure property relationships. TransFoxMol incorporates a multi-scale 2D molecular environment into a graph neural network + Transformer module and uses prior chemical maps to obtain a more focused attention landscape compared to that obtained using existing approaches. Experimental results show that TransFoxMol achieves state-of-the-art performance on MoleculeNet benchmarks and surpasses the performance of baselines that use self-supervised learning or geometry-enhanced strategies on small-scale datasets. Subsequent analyses indicate that TransFoxMol's predictions are highly interpretable and the clever use of chemical knowledge enables AI to perceive molecules in a simple but rational way, enhancing performance.

The advent of single-cell RNA sequencing (scRNA-seq) technologies has enabled gene expression profiling at the single-cell resolution, thereby enabling the quantification and comparison of transcriptional variability among individual cells. Although alterations in transcriptional variability have been observed in various biological states, statistical methods for quantifying and testing differential variability between groups of cells are still lacking. To identify the best practices in differential variability analysis of single-cell gene expression data, we propose and compare 12 statistical pipelines using different combinations of methods for normalization, feature selection, dimensionality reduction and variability calculation. Using high-quality synthetic scRNA-seq datasets, we benchmarked the proposed pipelines and found that the most powerful and accurate pipeline performs simple library size normalization, retains all genes in analysis and uses denSNE-based distances to cluster medoids as the variability measure. By applying this pipeline to scRNA-seq datasets of COVID-19 and autism patients, we have identified cellular variability changes between patients with different severity status or between patients and healthy controls.

The latent features extracted from the multiple sequence alignments (MSAs) of homologous protein families are useful for identifying residue-residue contacts, predicting mutation effects, shaping protein evolution, etc. Over the past three decades, a growing body of supervised and unsupervised machine learning methods have been applied to this field, yielding fruitful results. Here, we propose a novel self-supervised model, called encoder-transformation layer-decoder (ETLD) architecture, capable of capturing protein sequence latent features directly from MSAs. Compared to the typical autoencoder model, ETLD introduces a transformation layer with the ability to learn inter-site couplings, which can be used to parse out the two-dimensional residue-residue contacts map after a simple mathematical derivation or an additional supervised neural network. ETLD retains the process of encoding and decoding sequences, and the predicted probabilities of amino acids at each site can be further used to construct the mutation landscapes for mutation effects prediction, outperforming advanced models such as GEMME, DeepSequence and EVmutation in general. Overall, ETLD is a highly interpretable unsupervised model with great potential for improvement and can be further combined with supervised methods for more extensive and accurate predictions.

Molecular property prediction (MPP) is a crucial and fundamental task for AI-aided drug discovery (AIDD). Recent studies have shown great promise of applying self-supervised learning (SSL) to producing molecular representations to cope with the widely-concerned data scarcity problem in AIDD. As some specific substructures of molecules play important roles in determining molecular properties, molecular representations learned by deep learning models are expected to attach more importance to such substructures implicitly or explicitly to achieve better predictive performance. However, few SSL pre-trained models for MPP in the literature have ever focused on such substructures. To challenge this situation, this paper presents a Chemistry-Aware Fragmentation for Effective MPP (CAFE-MPP in short) under the self-supervised contrastive learning framework. First, a novel fragment-based molecular graph (FMG) is designed to represent the topological relationship between chemistry-aware substructures that constitute a molecule. Then, with well-designed hard negative pairs, a is pre-trained on fragment-level by contrastive learning to extract representations for the nodes in FMGs. Finally, a Graphormer model is leveraged to produce molecular representations for MPP based on the embeddings of fragments. Experiments on 11 benchmark datasets show that the proposed CAFE-MPP method achieves state-of-the-art performance on 7 of the 11 datasets and the second-best performance on 3 datasets, compared with six remarkable self-supervised methods. Further investigations also demonstrate that CAFE-MPP can learn to embed molecules into representations implicitly containing the information of fragments highly correlated to molecular properties, and can alleviate the over-smoothing problem of graph neural networks.

Alignment-based RNA-seq quantification methods typically involve a time-consuming alignment process prior to estimating transcript abundances. In contrast, alignment-free RNA-seq quantification methods bypass this step, resulting in significant speed improvements. Existing alignment-free methods rely on the Expectation-Maximization (EM) algorithm for estimating transcript abundances. However, EM algorithms only guarantee locally optimal solutions, leaving room for further accuracy improvement by finding a globally optimal solution. In this study, we present TQSLE, the first alignment-free RNA-seq quantification method that provides a globally optimal solution for transcript abundances estimation. TQSLE adopts a two-step approach: first, it constructs a k-mer frequency matrix A for the reference transcriptome and a k-mer frequency vector b for the RNA-seq reads; then, it directly estimates transcript abundances by solving the linear equation ATAx = ATb. We evaluated the performance of TQSLE using simulated and real RNA-seq data sets and observed that, despite comparable speed to other alignment-free methods, TQSLE outperforms them in terms of accuracy. TQSLE is freely available at https://github.com/yhg926/TQSLE.

Genome assembly is a computational technique that involves piecing together deoxyribonucleic acid (DNA) fragments generated by sequencing technologies to create a comprehensive and precise representation of the entire genome. Generating a high-quality human reference genome is a crucial prerequisite for comprehending human biology, and it is also vital for downstream genomic variation analysis. Many efforts have been made over the past few decades to create a complete and gapless reference genome for humans by using a diverse range of advanced sequencing technologies. Several available tools are aimed at enhancing the quality of haploid and diploid human genome assemblies, which include contig assembly, polishing of contig errors, scaffolding and variant phasing. Selecting the appropriate tools and technologies remains a daunting task despite several studies have investigated the pros and cons of different assembly strategies. The goal of this paper was to benchmark various strategies for human genome assembly by combining sequencing technologies and tools on two publicly available samples (NA12878 and NA24385) from Genome in a Bottle. We then compared their performances in terms of continuity, accuracy, completeness, variant calling and phasing. We observed that PacBio HiFi long-reads are the optimal choice for generating an assembly with low base errors. On the other hand, we were able to produce the most continuous contigs with Oxford Nanopore long-reads, but they may require further polishing to improve on quality. We recommend using short-reads rather than long-reads themselves to improve the base accuracy of contigs from Oxford Nanopore long-reads. Hi-C is the best choice for chromosome-level scaffolding because it can capture the longest-range DNA connectedness compared to 10× linked-reads and Bionano optical maps. However, a combination of multiple technologies can be used to further improve the quality and completeness of genome assembly. For diploid assembly, hifiasm is the best tool for human diploid genome assembly using PacBio HiFi and Hi-C data. Looking to the future, we expect that further advancements in human diploid assemblers will leverage the power of PacBio HiFi reads and other technologies with long-range DNA connectedness to enable the generation of high-quality, chromosome-level and haplotype-resolved human genome assemblies.

The availability of high-throughput sequencing data creates opportunities to comprehensively understand human diseases as well as challenges to train machine learning models using such high dimensions of data. Here, we propose a denoised multi-omics integration framework, which contains a distribution-based feature denoising algorithm, Feature Selection with Distribution (FSD), for dimension reduction and a multi-omics integration framework, Attention Multi-Omics Integration (AttentionMOI) to predict cancer prognosis and identify cancer subtypes. We demonstrated that FSD improved model performance either using single omic data or multi-omics data in 15 The Cancer Genome Atlas Program (TCGA) cancers for survival prediction and kidney cancer subtype identification. And our integration framework AttentionMOI outperformed machine learning models and current multi-omics integration algorithms with high dimensions of features. Furthermore, FSD identified features that were associated to cancer prognosis and could be considered as biomarkers.

Omics data from clinical samples are the predominant source of target discovery and drug development. Typically, hundreds or thousands of differentially expressed genes or proteins can be identified from omics data. This scale of possibilities is overwhelming for target discovery and validation using biochemical or cellular experiments. Most of these proteins and genes have no corresponding drugs or even active compounds. Moreover, a proportion of them may have been previously reported as being relevant to the disease of interest. To facilitate translational drug discovery from omics data, we have developed a new classification tool named Omics and Text driven Translational Medicine (OTTM). This tool can markedly narrow the range of proteins or genes that merit further validation via drug availability assessment and literature mining. For the 4489 candidate proteins identified in our previous proteomics study, OTTM recommended 40 FDA-approved or clinical trial drugs. Of these, 15 are available commercially and were tested on hepatocellular carcinoma Hep-G2 cells. Two drugs-tafenoquine succinate (an FDA-approved antimalarial drug targeting CYC1) and branaplam (a Phase 3 clinical drug targeting SMN1 for the treatment of spinal muscular atrophy)-showed potent inhibitory activity against Hep-G2 cell viability, suggesting that CYC1 and SMN1 may be potential therapeutic target proteins for hepatocellular carcinoma. In summary, OTTM is an efficient classification tool that can accelerate the discovery of effective drugs and targets using thousands of candidate proteins identified from omics data. The online and local versions of OTTM are available at http://otter-simm.com/ottm.html.

Accurate prediction of molecular properties is an important topic in drug discovery. Recent works have developed various representation schemes for molecular structures to capture different chemical information in molecules. The atom and motif can be viewed as hierarchical molecular structures that are widely used for learning molecular representations to predict chemical properties. Previous works have attempted to exploit both atom and motif to address the problem of information loss in single representation learning for various tasks. To further fuse such hierarchical information, the correspondence between learned chemical features from different molecular structures should be considered. Herein, we propose a novel framework for molecular property prediction, called hierarchical molecular graph neural networks (HimGNN). HimGNN learns hierarchical topology representations by applying graph neural networks on atom- and motif-based graphs. In order to boost the representational power of the motif feature, we design a Transformer-based local augmentation module to enrich motif features by introducing heterogeneous atom information in motif representation learning. Besides, we focus on the molecular hierarchical relationship and propose a simple yet effective rescaling module, called contextual self-rescaling, that adaptively recalibrates molecular representations by explicitly modelling interdependencies between atom and motif features. Extensive computational experiments demonstrate that HimGNN can achieve promising performances over state-of-the-art baselines on both classification and regression tasks in molecular property prediction.

Transmembrane proteins are receptors, enzymes, transporters and ion channels that are instrumental in regulating a variety of cellular activities, such as signal transduction and cell communication. Despite tremendous progress in computational capacities to support protein research, there is still a significant gap in the availability of specialized computational analysis toolkits for transmembrane protein research. Here, we introduce TMKit, an open-source Python programming interface that is modular, scalable and specifically designed for processing transmembrane protein data. TMKit is a one-stop computational analysis tool for transmembrane proteins, enabling users to perform database wrangling, engineer features at the mutational, domain and topological levels, and visualize protein-protein interaction interfaces. In addition, TMKit includes seqNetRR, a high-performance computing library that allows customized construction of a large number of residue connections. This library is particularly well suited for assigning correlation matrix-based features at a fast speed. TMKit should serve as a useful tool for researchers in assisting the study of transmembrane protein sequences and structures. TMKit is publicly available through https://github.com/2003100127/tmkit and https://tmkit-guide.herokuapp.com/doc/overview.

Structural variants (SVs) are genomic rearrangements that can take many different forms such as copy number alterations, inversions and translocations. During cell development and aging, somatic SVs accumulate in the genome with potentially neutral, deleterious or pathological effects. Generation of somatic SVs is a key mutational process in cancer development and progression. Despite their importance, the detection of somatic SVs is challenging, making them less studied than somatic single-nucleotide variants. In this review, we summarize recent advances in whole-genome sequencing (WGS)-based approaches for detecting somatic SVs at the tissue and single-cell levels and discuss their advantages and limitations. First, we describe the state-of-the-art computational algorithms for somatic SV calling using bulk WGS data and compare the performance of somatic SV detectors in the presence or absence of a matched-normal control. We then discuss the unique features of cutting-edge single-cell-based techniques for analyzing somatic SVs. The advantages and disadvantages of bulk and single-cell approaches are highlighted, along with a discussion of their sensitivity to copy-neutral SVs, usefulness for functional inferences and experimental and computational costs. Finally, computational approaches for linking somatic SVs to their functional readouts, such as those obtained from single-cell transcriptome and epigenome analyses, are illustrated, with a discussion of the promise of these approaches in health and diseases.

Precision medicine relies on the identification of robust disease and risk factor signatures from omics data. However, current knowledge-driven approaches may overlook novel or unexpected phenomena due to the inherent biases in biological knowledge. In this study, we present a data-driven signature discovery workflow for DNA methylation analysis utilizing network-coherent autoencoders (NCAEs) with biologically relevant latent embeddings. First, we explored the architecture space of autoencoders trained on a large-scale pan-tissue compendium (n = 75 272) of human epigenome-wide association studies. We observed the emergence of co-localized patterns in the deep autoencoder latent space representations that corresponded to biological network modules. We determined the NCAE configuration with the strongest co-localization and centrality signals in the human protein interactome. Leveraging the NCAE embeddings, we then trained interpretable deep neural networks for risk factor (aging, smoking) and disease (systemic lupus erythematosus) prediction and classification tasks. Remarkably, our NCAE embedding-based models outperformed existing predictors, revealing novel DNA methylation signatures enriched in gene sets and pathways associated with the studied condition in each case. Our data-driven biomarker discovery workflow provides a generally applicable pipeline to capture relevant risk factor and disease information. By surpassing the limitations of knowledge-driven methods, our approach enhances the understanding of complex epigenetic processes, facilitating the development of more effective diagnostic and therapeutic strategies.

Recent studies have demonstrated the significant role that circRNA plays in the progression of human diseases. Identifying circRNA-disease associations (CDA) in an efficient manner can offer crucial insights into disease diagnosis. While traditional biological experiments can be time-consuming and labor-intensive, computational methods have emerged as a viable alternative in recent years. However, these methods are often limited by data sparsity and their inability to explore high-order information. In this paper, we introduce a novel method named Knowledge Graph Encoder from Transformer for predicting CDA (KGETCDA). Specifically, KGETCDA first integrates more than 10 databases to construct a large heterogeneous non-coding RNA dataset, which contains multiple relationships between circRNA, miRNA, lncRNA and disease. Then, a biological knowledge graph is created based on this dataset and Transformer-based knowledge representation learning and attentive propagation layers are applied to obtain high-quality embeddings with accurately captured high-order interaction information. Finally, multilayer perceptron is utilized to predict the matching scores of CDA based on their embeddings. Our empirical results demonstrate that KGETCDA significantly outperforms other state-of-the-art models. To enhance user experience, we have developed an interactive web-based platform named HNRBase that allows users to visualize, download data and make predictions using KGETCDA with ease. The code and datasets are publicly available at https://github.com/jinyangwu/KGETCDA.

Genomic variants affecting pre-messenger RNA splicing and its regulation are known to underlie many rare genetic diseases. However, common workflows for genetic diagnosis and clinical variant interpretation frequently overlook splice-altering variants. To better serve patient populations and advance biomedical knowledge, it has become increasingly important to develop and refine approaches for detecting and interpreting pathogenic splicing variants. In this review, we will summarize a few recent developments and challenges in using RNA sequencing technologies for rare disease investigation. Moreover, we will discuss how recent computational splicing prediction tools have emerged as complementary approaches for revealing disease-causing variants underlying splicing defects. We speculate that continuous improvements to sequencing technologies and predictive modeling will not only expand our understanding of splicing regulation but also bring us closer to filling the diagnostic gap for rare disease patients.

Single-cell RNA sequencing (scRNA-seq) has emerged as a powerful tool for uncovering cellular heterogeneity. However, the high costs associated with this technique have rendered it impractical for studying large patient cohorts. We introduce ENIGMA (Deconvolution based on Regularized Matrix Completion), a method that addresses this limitation through accurately deconvoluting bulk tissue RNA-seq data into a readout with cell-type resolution by leveraging information from scRNA-seq data. By employing a matrix completion strategy, ENIGMA minimizes the distance between the mixture transcriptome obtained with bulk sequencing and a weighted combination of cell-type-specific expression. This allows the quantification of cell-type proportions and reconstruction of cell-type-specific transcriptomes. To validate its performance, ENIGMA was tested on both simulated and real datasets, including disease-related tissues, demonstrating its ability in uncovering novel biological insights.

MOTIVATION: Multiple displacement amplification (MDA) has become the most commonly used method of whole genome amplification, generating a vast amount of DNA with higher molecular weight and greater genome coverage. Coupling with long-read sequencing, it is possible to sequence the amplicons of over 20 kb in length. However, the formation of chimeric sequences (chimeras, expressed as structural errors in sequencing data) in MDA seriously interferes with the bioinformatics analysis but its influence on long-read sequencing data is unknown. RESULTS: We sequenced the phi29 DNA polymerase-mediated MDA amplicons on the PacBio platform and analyzed chimeras within the generated data. The 3rd-ChimeraMiner has been constructed as a pipeline for recognizing and restoring chimeras into the original structures in long-read sequencing data, improving the efficiency of using TGS data. Five long-read datasets and one high-fidelity long-read dataset with various amplification folds were analyzed. The result reveals that the mis-priming events in amplification are more frequently occurring than widely perceived, and the propor tion gradually accumulates from 42% to over 78% as the amplification continues. In total, 99.92% of recognized chimeric sequences were demonstrated to be artifacts, whose structures were wrongly formed in MDA instead of existing in original genomes. By restoring chimeras to their original structures, the vast majority of supplementary alignments that introduce false-positive structural variants are recycled, removing 97% of inversions on average and contributing to the analysis of structural variation in MDA-amplified samples. The impact of chimeras in long-read sequencing data analysis should be emphasized, and the 3rd-ChimeraMiner can help to quantify and reduce the influence of chimeras. AVAILABILITY AND IMPLEMENTATION: The 3rd-ChimeraMiner is available on GitHub, https://github.com/dulunar/3rdChimeraMiner.

MOTIVATION: Identifying the relationships among long non-coding RNAs (lncRNAs), microRNAs (miRNAs) and diseases is highly valuable for diagnosing, preventing, treating and prognosing diseases. The development of effective computational prediction methods can reduce experimental costs. While numerous methods have been proposed, they often to treat the prediction of lncRNA-disease associations (LDAs), miRNA-disease associations (MDAs) and lncRNA-miRNA interactions (LMIs) as separate task. Models capable of predicting all three relationships simultaneously remain relatively scarce. Our aim is to perform multi-task predictions, which not only construct a unified framework, but also facilitate mutual complementarity of information among lncRNAs, miRNAs and diseases. RESULTS: In this work, we propose a novel unsupervised embedding method called graph contrastive learning for multi-task prediction (GCLMTP). Our approach aims to predict LDAs, MDAs and LMIs by simultaneously extracting embedding representations of lncRNAs, miRNAs and diseases. To achieve this, we first construct a triple-layer lncRNA-miRNA-disease heterogeneous graph (LMDHG) that integrates the complex relationships between these entities based on their similarities and correlations. Next, we employ an unsupervised embedding model based on graph contrastive learning to extract potential topological feature of lncRNAs, miRNAs and diseases from the LMDHG. The graph contrastive learning leverages graph convolutional network architectures to maximize the mutual information between patch representations and corresponding high-level summaries of the LMDHG. Subsequently, for the three prediction tasks, multiple classifiers are explored to predict LDA, MDA and LMI scores. Comprehensive experiments are conducted on two datasets (from older and newer versions of the database, respectively). The results show that GCLMTP outperforms other state-of-the-art methods for the disease-related lncRNA and miRNA prediction tasks. Additionally, case studies on two datasets further demonstrate the ability of GCLMTP to accurately discover new associations. To ensure reproducibility of this work, we have made the datasets and source code publicly available at https://github.com/sheng-n/GCLMTP.

Adequate reporting is essential for evaluating the performance and clinical utility of a prognostic prediction model. Previous studies indicated a prevalence of incomplete or suboptimal reporting in translational and clinical studies involving development of multivariable prediction models for prognosis, which limited the potential applications of these models. While reporting templates introduced by the established guidelines provide an invaluable framework for reporting prognostic studies uniformly, there is a widespread lack of qualified adherence, which may be due to miscellaneous challenges in manual reporting of extensive model details, especially in the era of precision medicine. Here, we present ReProMSig (Reproducible Prognosis Molecular Signature), a web-based integrative platform providing the analysis framework for development, validation and application of a multivariable prediction model for cancer prognosis, using clinicopathological features and/or molecular profiles. ReProMSig platform supports transparent reporting by presenting both methodology details and analysis results in a strictly structured reporting file, following the guideline checklist with minimal manual input needed. The generated reporting file can be published together with a developed prediction model, to allow thorough interrogation and external validation, as well as online application for prospective cases. We demonstrated the utilities of ReProMSig by development of prognostic molecular signatures for stage II and III colorectal cancer respectively, in comparison with a published signature reproduced by ReProMSig. Together, ReProMSig provides an integrated framework for development, evaluation and application of prognostic/predictive biomarkers for cancer in a more transparent and reproducible way, which would be a useful resource for health care professionals and biomedical researchers.