BACKGROUND: Rapidly growing genome-wide ChIP-seq data have provided unprecedented opportunities to explore transcription factor (TF) binding under various cellular conditions. Despite the rich resources, development of analytical methods for studying the interaction among TFs in gene regulation still lags behind. RESULTS: In order to address cooperative TF binding and detect TF clusters with coordinative functions, we have developed novel computational methods based on clustering the sample paths of nonhomogeneous Poisson processes. Simulation studies demonstrated the capability of these methods to accurately detect TF clusters and uncover the hierarchy of TF interactions. A further application to the multiple-TF ChIP-seq data in mouse embryonic stem cells (ESCs) showed that our methods identified the cluster of core ESC regulators reported in the literature and provided new insights on functional implications of transcrisptional regulatory modules. CONCLUSIONS: Effective analytical tools are essential for studying protein-DNA relations. Information derived from this research will help us better understand the orchestration of transcription factors in gene regulation processes.

BACKGROUND: The central role of proteins in diseases has made them increasingly attractive as therapeutic targets and indicators of cellular processes. Protein microarrays are emerging as an important means of characterising protein activity. Their accurate downstream analysis to produce biologically significant conclusions is largely dependent on proper pre-processing of extracted signal intensities. However, existing computational tools are not specifically tailored to the nature of these data and lack unanimity. RESULTS: Here, we present the single-channel Protein Microarray Analysis Pipeline, a tailored computational tool for analysis of single-channel protein microarrays enabling biomarker identification, implemented in R, and as an interactive web application. We compared four existing background correction and normalization methods as well as three array filtering techniques, applied to four real datasets with two microarray designs, extracted using two software programs. The normexp, cyclic loess, and array weighting methods were most effective for background correction, normalization, and filtering respectively. CONCLUSIONS: Thus, here we provided a versatile and effective pre-processing and differential analysis workflow for single-channel protein microarray data in form of an R script and web application ( https://metaomics.uct.ac.za/shinyapps/Pro-MAP/ .) for those not well versed in the R programming language.

BACKGROUND: Knowing the responses of a patient to drugs is essential to make personalized medicine practical. Since the current clinical drug response experiments are time-consuming and expensive, utilizing human genomic information and drug molecular characteristics to predict drug responses is of urgent importance. Although a variety of computational drug response prediction methods have been proposed, their effectiveness is still not satisfying. RESULTS: In this study, we propose a method called LGRDRP (Learning Graph Representation for Drug Response Prediction) to predict cell line-drug responses. At first, LGRDRP constructs a heterogeneous network integrating multiple kinds of information: cell line miRNA expression profiles, drug chemical structure similarity, gene-gene interaction, cell line-gene interaction and known cell line-drug responses. Then, for each cell line, learning graph representation and Laplacian feature selection are combined to obtain network topology features related to the cell line. The learning graph representation method learns network topology structure features, and the Laplacian feature selection method further selects out some most important ones from them. Finally, LGRDRP trains an SVM model to predict drug responses based on the selected features of the known cell line-drug responses. Our five-fold cross-validation results show that LGRDRP is significantly superior to the art-of-the-state methods in the measures of the average area under the receiver operating characteristics curve, the average area under the precision-recall curve and the recall rate of top-k predicted sensitive cell lines. CONCLUSIONS: Our results demonstrated that the usage of multiple types of information about cell lines and drugs, the learning graph representation method, and the Laplacian feature selection is useful to the improvement of performance in predicting drug responses. We believe that such an approach would be easily extended to similar problems such as miRNA-disease relationship inference.

BACKGROUND: Modern biomedical research is data-driven and relies heavily on the re-use and sharing of data. Biomedical data, however, is subject to strict data protection requirements. Due to the complexity of the data required and the scale of data use, obtaining informed consent is often infeasible. Other methods, such as anonymization or federation, in turn have their own limitations. Secure multi-party computation (SMPC) is a cryptographic technology for distributed calculations, which brings formally provable security and privacy guarantees and can be used to implement a wide-range of analytical approaches. As a relatively new technology, SMPC is still rarely used in real-world biomedical data sharing activities due to several barriers, including its technical complexity and lack of usability. RESULTS: To overcome these barriers, we have developed the tool EasySMPC, which is implemented in Java as a cross-platform, stand-alone desktop application provided as open-source software. The tool makes use of the SMPC method Arithmetic Secret Sharing, which allows to securely sum up pre-defined sets of variables among different parties in two rounds of communication (input sharing and output reconstruction) and integrates this method into a graphical user interface. No additional software services need to be set up or configured, as EasySMPC uses the most widespread digital communication channel available: e-mails. No cryptographic keys need to be exchanged between the parties and e-mails are exchanged automatically by the software. To demonstrate the practicability of our solution, we evaluated its performance in a wide range of data sharing scenarios. The results of our evaluation show that our approach is scalable (summing up 10,000 variables between 20 parties takes less than 300 s) and that the number of participants is the essential factor. CONCLUSIONS: We have developed an easy-to-use "no-code solution" for performing secure joint calculations on biomedical data using SMPC protocols, which is suitable for use by scientists without IT expertise and which has no special infrastructure requirements. We believe that innovative approaches to data sharing with SMPC are needed to foster the translation of complex protocols into practice.

BACKGROUND: Recent increasing evidence indicates that three-dimensional chromosome structure plays an important role in genomic function. Topologically associating domains (TADs) are self-interacting regions that have been shown to be a chromosomal structural unit. During evolution, these are conserved based on checking synteny block cross species. Are there common TAD patterns across species or cell lines? RESULTS: To address the above question, we propose a novel task-TAD recognition-as opposed to traditional TAD identification. Specifically, we treat Hi-C maps as images, thus re-casting TAD recognition as image pattern recognition, for which we use a convolutional neural network and a residual neural network. In addition, we propose an elegant way to generate non-TAD data for binary classification. We demonstrate deep learning performance which is quite promising, AUC > 0.80, through cross-species and cell-type validation. CONCLUSIONS: TADs have been shown to be conserved during evolution. Interestingly, our results confirm that the TAD recognition model is practical across species, which indicates that TADs between human and mouse show common patterns from an image classification point of view. Our approach could be a new way to identify TAD variations or patterns among Hi-C maps. For example, TADs of two Hi-C maps are conserved if the two classification models are exchangeable.

BACKGROUND: Single-cell RNA-sequencing (scRNA-seq) technologies allow for the study of gene expression in individual cells. Often, it is of interest to understand how transcriptional activity is associated with cell-specific covariates, such as cell type, genotype, or measures of cell health. Traditional approaches for this type of association mapping assume independence between the outcome variables (or genes), and perform a separate regression for each. However, these methods are computationally costly and ignore the substantial correlation structure of gene expression. Furthermore, count-based scRNA-seq data pose challenges for traditional models based on Gaussian assumptions. RESULTS: We aim to resolve these issues by developing a reduced-rank regression model that identifies low-dimensional linear associations between a large number of cell-specific covariates and high-dimensional gene expression readouts. Our probabilistic model uses a Poisson likelihood in order to account for the unique structure of scRNA-seq counts. We demonstrate the performance of our model using simulations, and we apply our model to a scRNA-seq dataset, a spatial gene expression dataset, and a bulk RNA-seq dataset to show its behavior in three distinct analyses. CONCLUSION: We show that our statistical modeling approach, which is based on reduced-rank regression, captures associations between gene expression and cell- and sample-specific covariates by leveraging low-dimensional representations of transcriptional states.

BACKGROUND: Localization-based super-resolution microscopy resolves macromolecular structures down to a few nanometers by computationally reconstructing fluorescent emitter coordinates from diffraction-limited spots. The most commonly used algorithms are based on fitting parametric models of the point spread function (PSF) to a measured photon distribution. These algorithms make assumptions about the symmetry of the PSF and thus, do not work well with irregular, non-linear PSFs that occur for example in confocal lifetime imaging, where a laser is scanned across the sample. An alternative method for reconstructing sparse emitter sets from noisy, diffraction-limited images is compressed sensing, but due to its high computational cost it has not yet been widely adopted. Deep neural network fitters have recently emerged as a new competitive method for localization microscopy. They can learn to fit arbitrary PSFs, but require extensive simulated training data and do not generalize well. A method to efficiently fit the irregular PSFs from confocal lifetime localization microscopy combining the advantages of deep learning and compressed sensing would greatly improve the acquisition speed and throughput of this method. RESULTS: Here we introduce ReCSAI, a compressed sensing neural network to reconstruct localizations for confocal dSTORM, together with a simulation tool to generate training data. We implemented and compared different artificial network architectures, aiming to combine the advantages of compressed sensing and deep learning. We found that a U-Net with a recursive structure inspired by iterative compressed sensing showed the best results on realistic simulated datasets with noise, as well as on real experimentally measured confocal lifetime scanning data. Adding a trainable wavelet denoising layer as prior step further improved the reconstruction quality. CONCLUSIONS: Our deep learning approach can reach a similar reconstruction accuracy for confocal dSTORM as frame binning with traditional fitting without requiring the acquisition of multiple frames. In addition, our work offers generic insights on the reconstruction of sparse measurements from noisy experimental data by combining compressed sensing and deep learning. We provide the trained networks, the code for network training and inference as well as the simulation tool as python code and Jupyter notebooks for easy reproducibility.

BACKGROUND: The application of PacBio HiFi and ultra-long ONT reads have enabled huge progress in the contig-level assembly, but it is still challenging to assemble large contigs into chromosomes with available Hi-C scaffolding tools, which count Hi-C links between contigs using the whole or a large part of contig regions. As the Hi-C links of two adjacent contigs concentrate only at the neighbor ends of the contigs, larger contig size will reduce the power to differentiate adjacent (signal) and non-adjacent (noise) contig linkages, leading to a higher rate of mis-assembly. RESULTS: We design and develop a novel Hi-C based scaffolding tool EndHiC, which is suitable to assemble large contigs into chromosomal-level scaffolds. The core idea behind EndHiC, which distinguishes it from other Hi-C scaffolding tools, is using Hi-C links only from the most effective regions of contig ends. By this way, the signal neighbor contig linkages and noise non-neighbor contig linkages are separated more clearly. Benefiting from the increased signal to noise ratio, the reciprocal best requirement, as well as the robustness evaluation, EndHiC achieves higher accuracy for scaffolding large contigs compared to existing tools. EndHiC has been successfully applied in the Hi-C scaffolding of simulated data from human, rice and Arabidopsis, and real data from human, great burdock, water spinach, chicory, endive, yacon, and Ipomoea cairica, suggesting that EndHiC can be applied to a broad range of plant and animal genomes. CONCLUSIONS: EndHiC is a novel Hi-C scaffolding tool, which is suitable for scaffolding of contig assemblies with contig N50 size near or over 10 Mb and N90 size near or over 1 Mb. EndHiC is efficient both in time and memory, and it is interface-friendly to the users. As more genome projects have been launched and the contig continuity constantly improved, we believe EndHiC has the potential to make a great contribution to the genomics field and liberate the scientists from labor-intensive manual curation works.

BACKGROUND: Researchers have tried to identify and count different blood cells in microscopic smear images by using deep learning methods of artificial intelligence to solve the highly time-consuming problem. RESULTS: The three types of blood cells are platelets, red blood cells, and white blood cells. This study used the Resnet50 network as a backbone network of the single shot detector (SSD) for automatically identifying and counting different blood cells and, meanwhile, proposed a systematic method to find a better combination of algorithm hyperparameters of the Resnet50 network for promoting accuracy for identifying and counting blood cells. The Resnet50 backbone network of the SSD with its optimized algorithm hyperparameters, which is called the Resnet50-SSD model, was developed to enhance the feature extraction ability for identifying and counting blood cells. Furthermore, the algorithm hyperparameters of Resnet50 backbone networks of the SSD were optimized by the Taguchi experimental method for promoting detection accuracy of the Resnet50-SSD model. The experimental result shows that the detection accuracy of the Resnet50-SSD model with 512 × 512 × 3 input images was better than that of the Resnet50-SSD model with 300 × 300 × 3 input images on the test set of blood cells images. Additionally, the detection accuracy of the Resnet50-SSD model using the combination of algorithm hyperparameters got by the Taguchi method was better than that of the Resnet50-SSD model using the combination of algorithm hyperparameters given by the Matlab example. CONCLUSION: In blood cell images acquired from the BCCD dataset, the proposed Resnet50-SSD model had higher accuracy in identifying and counting blood cells, especially white blood cells and red blood cells.

BACKGROUND: Drug-drug interactions (DDIs) are preventable causes of medical injuries and often result in doctor and emergency room visits. Previous research demonstrates the effectiveness of using matrix completion approaches based on known drug interactions to predict unknown Drug-drug interactions. However, in the case of a new drug, where there is limited or no knowledge regarding the drug's existing interactions, such an approach is unsuitable, and other drug's preferences can be used to accurately predict new Drug-drug interactions. METHODS: We propose adjacency biomedical text embedding (ABTE) to address this limitation by using a hybrid approach which combines known drugs' interactions and the drug's biomedical text embeddings to predict the DDIs of both new and well known drugs. RESULTS: Our evaluation demonstrates the superiority of this approach compared to recently published DDI prediction models and matrix factorization-based approaches. Furthermore, we compared the use of different text embedding methods in ABTE, and found that the concept embedding approach, which involves biomedical information in the embedding process, provides the highest performance for this task. Additionally, we demonstrate the effectiveness of leveraging biomedical text embedding for additional drugs' biomedical prediction task by presenting text embedding's contribution to a multi-modal pregnancy drug safety classification. CONCLUSION: Text and concept embeddings created by analyzing a domain-specific large-scale biomedical corpora can be used for predicting drug-related properties such as Drug-drug interactions and drug safety prediction. Prediction models based on the embeddings resulted in comparable results to hand-crafted features, however text embeddings do not require manual categorization or data collection and rely solely on the published literature.

Preeclampsia (PE) has an increasing incidence worldwide, and there is no gold standard for prediction. Recent progress has shown that abnormal decidualization and impaired vascular remodeling are essential to PE pathogenesis. Therefore, it is of great significance to analyze the decidua basalis and blood changes of PE to explore new methods. Here, we performed weighted gene co-expression network analysis based on 9553 differentially expressed genes of decidua basalis data (GSE60438 includes 25 cases of PE and 23 non-cases) from Gene Expression Omnibus to screen relevant module-eigengenes (MEs). Among them, MEblue and MEgrey are the most correlated with PE, which contains 371 core genes. Subsequently, we applied the logistic least absolute shrinkage and selection operator regression, screened 43 genes most relevant to prediction from the intersections of the 371 genes and training set (GSE48424 includes 18 cases of PE and 18 non-cases) genes, and built a predictive model. The specificity and sensitivity are illustrated by receiver operating characteristic curves, and the stability was verified by two validation sets (GSE86200 includes 12 cases of PE and 48 non-cases, and GSE85307 includes 47 cases of PE and 110 non-cases). The results demonstrated that our predictive model shows good predictions, with an area under the curve of 0.991 for the training set, 0.874 and 0.986 for the validation sets. Finally, we found the 43 key marker genes in the model are closely associated with the clinically accepted predictive molecules, including FLT1, PIGF, ENG and VEGF. Therefore, this predictive model provides a potential approach for PE diagnosis and treatment.

BACKGROUND: The roles of antibody and antigen are indispensable in targeted diagnosis, therapy, and biomedical discovery. On top of that, massive numbers of new scientific articles about antibodies and/or antigens are published each year, which is a precious knowledge resource but has yet been exploited to its full potential. We, therefore, aim to develop a biomedical natural language processing tool that can automatically identify antibody and antigen entities from articles. RESULTS: We first annotated an antibody-antigen corpus including 3210 relevant PubMed abstracts using a semi-automatic approach. The Inter-Annotator Agreement score of 3 annotators ranges from 91.46 to 94.31%, indicating that the annotations are consistent and the corpus is reliable. We then used the corpus to develop and optimize BiLSTM-CRF-based and BioBERT-based models. The models achieved overall F1 scores of 62.49% and 81.44%, respectively, which showed potential for newly studied entities. The two models served as foundation for development of a named entity recognition (NER) tool that automatically recognizes antibody and antigen names from biomedical literature. CONCLUSIONS: Our antibody-antigen NER models enable users to automatically extract antibody and antigen names from scientific articles without manually scanning through vast amounts of data and information in the literature. The output of NER can be used to automatically populate antibody-antigen databases, support antibody validation, and facilitate researchers with the most appropriate antibodies of interest. The packaged NER model is available at https://github.com/TrangDinh44/ABAG_BioBERT.git .

BACKGROUND: The correct establishment of the barcode classification system for fish can facilitate biotaxonomists to distinguish fish species, and it can help the government to verify the authenticity of the ingredients of fish products or identify unknown fish related samples. The Cytochrome c oxidation I (COI) gene sequence in the mitochondria of each species possesses unique characteristics, which has been widely used as barcodes in identifying species in recent years. Instead of using COI gene sequences for primer design, flanking tRNA segments of COI genes from 2618 complete fish mitochondrial genomes were analyzed to discover suitable primers for fish classification at taxonomic family level. The minimal number of primer sets is designed to effectively distinguish various clustered groups of fish species for identification applications. Sequence alignment analysis and cross tRNA segment comparisons were applied to check and ensure the primers for each cluster group are exclusive. RESULTS: Two approaches were applied to improve primer design and re-cluster fish species. The results have shown that exclusive primers for 2618 fish species were successfully discovered through in silico analysis. In addition, we applied sequence alignment analysis to confirm that each pair of primers can successfully identify all collected fish species at the taxonomic family levels. CONCLUSIONS: This study provided a practical strategy to discover unique primers for each fishery species and a comprehensive list of exclusive primers for extracting COI barcode sequences of all known fishery species. Various applications of verification of fish products or identification of unknown fish species could be effectively achieved.

BACKGROUND: A deep understanding of carcinogenesis at the DNA level underpins many advances in cancer prevention and treatment. Mutational signatures provide a breakthrough conceptualisation, as well as an analysis framework, that can be used to build such understanding. They capture somatic mutation patterns and at best identify their causes. Most studies in this context have focused on an inherently additive analysis, e.g. by non-negative matrix factorization, where the mutations within a cancer sample are explained by a linear combination of independent mutational signatures. However, other recent studies show that the mutational signatures exhibit non-additive interactions. RESULTS: We carefully analysed such additive model fits from the PCAWG study cataloguing mutational signatures as well as their activities across thousands of cancers. Our analysis identified systematic and non-random structure of residuals that is left unexplained by the additive model. We used hierarchical clustering to identify cancer subsets with similar residual profiles to show that both systematic mutation count overestimation and underestimation take place. We propose an extension to the additive mutational signature model-multiplicatively acting modulatory processes-and develop a maximum-likelihood framework to identify such modulatory mutational signatures. The augmented model is expressive enough to almost fully remove the observed systematic residual patterns. CONCLUSION: We suggest the modulatory processes biologically relate to sample specific DNA repair propensities with cancer or tissue type specific profiles. Overall, our results identify an interesting direction where to expand signature analysis.

Accurate estimate of relatedness is important for genetic data analyses, such as heritability estimation and association mapping based on data collected from genome-wide association studies. Inaccurate relatedness estimates may lead to biased heritability estimations and spurious associations. Individual-level genotype data are often used to estimate kinship coefficient between individuals. The commonly used sample correlation-based genomic relationship matrix (scGRM) method estimates kinship coefficient by calculating the average sample correlation coefficient among all single nucleotide polymorphisms (SNPs), where the observed allele frequencies are used to calculate both the expectations and variances of genotypes. Although this method is widely used, a substantial proportion of estimated kinship coefficients are negative, which are difficult to interpret. In this paper, through mathematical derivation, we show that there indeed exists bias in the estimated kinship coefficient using the scGRM method when the observed allele frequencies are regarded as true frequencies. This leads to negative bias for the average estimate of kinship among all individuals, which explains the estimated negative kinship coefficients. Based on this observation, we propose an unbiased estimation method, UKin, which can reduce kinship estimation bias. We justify our improved method with rigorous mathematical proof. We have conducted simulations as well as two real data analyses to compare UKin with scGRM and three other kinship estimating methods: rGRM, tsGRM, and KING. Our results demonstrate that both bias and root mean square error in kinship coefficient estimation could be reduced by using UKin. We further investigated the performance of UKin, KING, and three GRM-based methods in calculating the SNP-based heritability, and show that UKin can improve estimation accuracy for heritability regardless of the scale of SNP panel.

BACKGROUND: Glaucoma can cause irreversible blindness to people's eyesight. Since there are no symptoms in its early stage, it is particularly important to accurately segment the optic disc (OD) and optic cup (OC) from fundus medical images for the screening and prevention of glaucoma. In recent years, the mainstream method of OD and OC segmentation is convolution neural network (CNN). However, most existing CNN methods segment OD and OC separately and ignore the a priori information that OC is always contained inside the OD region, which makes the segmentation accuracy of most methods not high enough. METHODS: This paper proposes a new encoder-decoder segmentation structure, called RSAP-Net, for joint segmentation of OD and OC. We first designed an efficient U-shaped segmentation network as the backbone. Considering the spatial overlap relationship between OD and OC, a new Residual spatial attention path is proposed to connect the encoder-decoder to retain more characteristic information. In order to further improve the segmentation performance, a pre-processing method called MSRCR-PT (Multi-Scale Retinex Colour Recovery and Polar Transformation) has been devised. It incorporates a multi-scale Retinex colour recovery algorithm and a polar coordinate transformation, which can help RSAP-Net to produce more refined boundaries of the optic disc and the optic cup. RESULTS: The experimental results show that our method achieves excellent segmentation performance on the Drishti-GS1 standard dataset. In the OD and OC segmentation effects, the F1 scores are 0.9752 and 0.9012, respectively. The BLE are 6.33 pixels and 11.97 pixels, respectively. CONCLUSIONS: This paper presents a new framework for the joint segmentation of optic discs and optic cups, called RSAP-Net. The framework mainly consists of a U-shaped segmentation skeleton and a residual space attention path module. The design of a pre-processing method called MSRCR-PT for the OD/OC segmentation task can improve segmentation performance. The method was evaluated on the publicly available Drishti-GS1 standard dataset and proved to be effective.

BACKGROUND: Independent Component Analysis (ICA) allows the dissection of omic datasets into modules that help to interpret global molecular signatures. The inherent randomness of this algorithm can be overcome by clustering many iterations of ICA together to obtain robust components. Existing algorithms for robust ICA are dependent on the choice of clustering method and on computing a potentially biased and large Pearson distance matrix. RESULTS: We present robustica, a Python-based package to compute robust independent components with a fully customizable clustering algorithm and distance metric. Here, we exploited its customizability to revisit and optimize robust ICA systematically. Of the 6 popular clustering algorithms considered, DBSCAN performed the best at clustering independent components across ICA iterations. To enable using Euclidean distances, we created a subroutine that infers and corrects the components' signs across ICA iterations. Our subroutine increased the resolution, robustness, and computational efficiency of the algorithm. Finally, we show the applicability of robustica by dissecting over 500 tumor samples from low-grade glioma (LGG) patients, where we define two new gene expression modules with key modulators of tumor progression upon IDH1 and TP53 mutagenesis. CONCLUSION: robustica brings precise, efficient, and customizable robust ICA into the Python toolbox. Through its customizability, we explored how different clustering algorithms and distance metrics can further optimize robust ICA. Then, we showcased how robustica can be used to discover gene modules associated with combinations of features of biological interest. Taken together, given the broad applicability of ICA for omic data analysis, we envision robustica will facilitate the seamless computation and integration of robust independent components in large pipelines.

BACKGROUND: Protein-protein interactions are widespread in biological systems and play an important role in cell biology. Since traditional laboratory-based methods have some drawbacks, such as time-consuming, money-consuming, etc., a large number of methods based on deep learning have emerged. However, these methods do not take into account the long-distance dependency information between each two amino acids in sequence. In addition, most existing models based on graph neural networks only aggregate the first-order neighbors in protein-protein interaction (PPI) network. Although multi-order neighbor information can be aggregated by increasing the number of layers of neural network, it is easy to cause over-fitting. So, it is necessary to design a network that can capture long distance dependency information between amino acids in the sequence and can directly capture multi-order neighbor information in protein-protein interaction network. RESULTS: In this study, we propose a multi-hop neural network (LDMGNN) model combining long distance dependency information to predict the multi-label protein-protein interactions. In the LDMGNN model, we design the protein amino acid sequence encoding (PAASE) module with the multi-head self-attention Transformer block to extract the features of amino acid sequences by calculating the interdependence between every two amino acids. And expand the receptive field in space by constructing a two-hop protein-protein interaction (THPPI) network. We combine PPI network and THPPI network with amino acid sequence features respectively, then input them into two identical GIN blocks at the same time to obtain two embeddings. Next, the two embeddings are fused and input to the classifier for predict multi-label protein-protein interactions. Compared with other state-of-the-art methods, LDMGNN shows the best performance on both the SHS27K and SHS148k datasets. Ablation experiments show that the PAASE module and the construction of THPPI network are feasible and effective. CONCLUSIONS: In general terms, our proposed LDMGNN model has achieved satisfactory results in the prediction of multi-label protein-protein interactions.

BACKGROUND: Monoclonal antibodies (mAbs) have been used as therapeutic agents, which must overcome many developability issues after the discovery from in vitro display libraries. Especially, polyreactive mAbs can strongly bind to a specific target and weakly bind to off-target proteins, which leads to poor antibody pharmacokinetics in clinical development. Although early assessment of polyreactive mAbs is important in the early discovery stage, experimental assessments are usually time-consuming and expensive. Therefore, computational approaches for predicting the polyreactivity of single-chain fragment variables (scFvs) in the early discovery stage would be promising for reducing experimental efforts. RESULTS: Here, we made prediction models for the polyreactivity of scFvs with the known polyreactive antibody features and natural language model descriptors. We predicted 19,426 protein structures of scFvs with trRosetta to calculate the polyreactive antibody features and investigated the classifying performance of each factor for polyreactivity. In the known polyreactive features, the net charge of the CDR2 loop, the tryptophan and glycine residues in CDR-H3, and the lengths of the CDR1 and CDR2 loops, importantly contributed to the performance of the models. Additionally, the hydrodynamic features, such as partial specific volume, gyration radius, and isoelectric points of CDR loops and scFvs, were newly added to improve model performance. Finally, we made the prediction model with a robust performance ([Formula: see text]) with an ensemble learning of the top 3 best models. CONCLUSION: The prediction models for polyreactivity would help assess polyreactive scFvs in the early discovery stage and our approaches would be promising to develop machine learning models with quantitative data from high throughput assays for antibody screening.

BACKGROUND: Gene set enrichment analysis (detecting phenotypic terms that emerge as significant in a set of genes) plays an important role in bioinformatics focused on diseases of genetic basis. To facilitate phenotype-oriented gene set analysis, we developed PhenoExam, a freely available R package for tool developers and a web interface for users, which performs: (1) phenotype and disease enrichment analysis on a gene set; (2) measures statistically significant phenotype similarities between gene sets and (3) detects significant differential phenotypes or disease terms across different databases. RESULTS: PhenoExam generates sensitive and accurate phenotype enrichment analyses. It is also effective in segregating gene sets or Mendelian diseases with very similar phenotypes. We tested the tool with two similar diseases (Parkinson and dystonia), to show phenotype-level similarities but also potentially interesting differences. Moreover, we used PhenoExam to validate computationally predicted new genes potentially associated with epilepsy. CONCLUSIONS: We developed PhenoExam, a freely available R package and Web application, which performs phenotype enrichment and disease enrichment analysis on gene set G, measures statistically significant phenotype similarities between pairs of gene sets G and G' and detects statistically significant exclusive phenotypes or disease terms, across different databases. We proved with simulations and real cases that it is useful to distinguish between gene sets or diseases with very similar phenotypes. Github R package URL is https://github.com/alexcis95/PhenoExam . Shiny App URL is https://alejandrocisterna.shinyapps.io/phenoexamweb/ .

There is evidence that DNA breathing (spontaneous opening of the DNA strands) plays a relevant role in the interactions of DNA with other molecules, and in particular in the transcription process. Therefore, having physical models that can predict these openings is of interest. However, this source of information has not been used before either in transcription start sites (TSSs) or promoter prediction. In this article, one such model is used as an additional information source that, when used by a machine learning (ML) model, improves the results of current methods for the prediction of TSSs. In addition, we provide evidence on the validity of the physical model, as it is able by itself to predict TSSs with high accuracy. This opens an exciting avenue of research at the intersection of statistical mechanics and ML, where ML models in bioinformatics can be improved using physical models of DNA as feature extractors.

BACKGROUND: Escherichia coli Nissle 1917 (EcN) is a probiotic bacterium used to treat various gastrointestinal diseases. EcN is increasingly being used as a chassis for the engineering of advanced microbiome therapeutics. To aid in future engineering efforts, our aim was to construct an updated metabolic model of EcN with extended secondary metabolite representation. RESULTS: An updated high-quality genome-scale metabolic model of EcN, iHM1533, was developed based on comparison with 55 E. coli/Shigella reference GEMs and manual curation, including expanded secondary metabolite pathways (enterobactin, salmochelins, aerobactin, yersiniabactin, and colibactin). The model was validated and improved using phenotype microarray data, resulting in an 82.3% accuracy in predicting growth phenotypes on various nutrition sources. Flux variability analysis with previously published 13C fluxomics data validated prediction of the internal central carbon fluxes. A standardised test suite called Memote assessed the quality of iHM1533 to have an overall score of 89%. The model was applied by using constraint-based flux analysis to predict targets for optimisation of secondary metabolite production. Modelling predicted design targets from across amino acid metabolism, carbon metabolism, and other subsystems that are common or unique for influencing the production of various secondary metabolites. CONCLUSION: iHM1533 represents a well-annotated metabolic model of EcN with extended secondary metabolite representation. Phenotype characterisation and the iHM1533 model provide a better understanding of the metabolic capabilities of EcN and will help future metabolic engineering efforts.

BACKGROUND: Identifying drug-target interactions (DTIs) plays a key role in drug development. Traditional wet experiments to identify DTIs are expensive and time consuming. Effective computational methods to predict DTIs are useful to narrow the searching scope of potential drugs and speed up the process of drug discovery. There are a variety of non-negativity matrix factorization based methods to predict DTIs, but the convergence of the algorithms used in the matrix factorization are often overlooked and the results can be further improved. RESULTS: In order to predict DTIs more accurately and quickly, we propose an alternating direction algorithm to solve graph regularized non-negative matrix factorization with prior knowledge consistency constraint (ADA-GRMFC). Based on known DTIs, drug chemical structures and target sequences, ADA-GRMFC at first constructs a DTI matrix, a drug similarity matrix and a target similarity matrix. Then DTI prediction is modeled as the non-negative factorization of the DTI matrix with graph dual regularization terms and a prior knowledge consistency constraint. The graph dual regularization terms are used to integrate the information from the drug similarity matrix and the target similarity matrix, and the prior knowledge consistency constraint is used to ensure the matrix decomposition result should be consistent with the prior knowledge of known DTIs. Finally, an alternating direction algorithm is used to solve the matrix factorization. Furthermore, we prove that the algorithm can converge to a stationary point. Extensive experimental results of 10-fold cross-validation show that ADA-GRMFC has better performance than other state-of-the-art methods. In the case study, ADA-GRMFC is also used to predict the targets interacting with the drug olanzapine, and all of the 10 highest-scoring targets have been accurately predicted. In predicting drug interactions with target estrogen receptors alpha, 17 of the 20 highest-scoring drugs have been validated.

BACKGROUND: CircRNAs are essential for the regulation of post-transcriptional gene expression, including as miRNA sponges, and play an important role in disease development. Some computational tools have been proposed recently to predict circRNA, since only one classifier is used, there is still much that can be done to improve the performance. RESULTS: StackCirRNAPred was proposed, the computational classification of long circRNA from other lncRNA based on stacking strategy. In order to cope with the potential problem that a single feature might not be able to distinguish circRNA well from other lncRNA, we first extracted features from different sources, including nucleic acid composition, sequence spatial features and physicochemical properties, Alu and tandem repeats. We innovatively apply the stacking strategy to integrate the more advantageous classifiers of RF, LightGBM, XGBoost. This allows the model to incorporate these features more flexibly. StackCirRNAPred was found to be significantly better than other tools, with precision, accuracy, F1, recall and MCC of 0.843, 0.833, 0.831, 0.819 and 0.666 respectively. We tested it directly on the mouse dataset. StackCirRNAPred was still significantly better than other methods, with precision, accuracy, F1, recall and MCC of 0.837, 0.839, 0.839, 0.841, 0.677. CONCLUSIONS: We proposed StackCirRNAPred based on stacking strategy to distinguish long circRNAs from other lncRNAs. With the test results demonstrating the validity and robustness of StackCirRNAPred, we hope StackCirRNAPred will complement existing circRNA prediction methods and is helpful in down-stream research.

BACKGROUND: Lung adenocarcinoma (LUAD) is the most common malignant tumor that seriously affects human health. Previous studies have indicated that abnormal levels of glycosylation promote progression and poor prognosis of lung cancer. Thus, the present study aimed to explore the prognostic signature related to glycosyltransferases (GTs) for LUAD. METHODS: The gene expression profiles were obtained from The Cancer Genome Atlas (TCGA) database, and GTs were obtained from the GlycomeDB database. Differentially expressed GTs-related genes (DGTs) were identified using edge package and Venn diagram. Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and ingenuity pathway analysis (IPA) methods were used to investigate the biological processes of DGTs. Subsequently, Cox and Least Absolute Shrinkage and Selection Operator (LASSO) regression analyses were performed to construct a prognostic model for LUAD. Kaplan-Meier (K-M) analysis was adopted to explore the overall survival (OS) of LUAD patients. The accuracy and specificity of the prognostic model were evaluated by receiver operating characteristic analysis (ROC). In addition, single-sample gene set enrichment analysis (ssGSEA) algorithm was used to analyze the infiltrating immune cells in the tumor environment. RESULTS: A total of 48 DGTs were mainly enriched in the processes of glycosylation, glycoprotein biosynthetic process, glycosphingolipid biosynthesis-lacto and neolacto series, and cell-mediated immune response. Furthermore, B3GNT3, MFNG, GYLTL1B, ALG3, and GALNT13 were screened as prognostic genes to construct a risk model for LUAD, and the LUAD patients were divided into high- and low-risk groups. K-M curve suggested that patients with a high-risk score had shorter OS than those with a low-risk score. The ROC analysis demonstrated that the risk model efficiently diagnoses LUAD. Additionally, the proportion of infiltrating aDCs (p < 0.05) and Tgds (p < 0.01) was higher in the high-risk group than in the low-risk group. Spearman's correlation analysis manifested that the prognostic genes (MFNG and ALG3) were significantly correlated with infiltrating immune cells. CONCLUSION: In summary, this study established a novel GTs-related risk model for the prognosis of LUAD patients, providing new therapeutic targets for LUAD. However, the biological role of glycosylation-related genes in LUAD needs to be explored further.

BACKGROUND: Drug-drug interactions (DDIs) occur when two or more drugs are taken simultaneously or successively. Early detection of adverse drug interactions can be essential in preventing medical errors and reducing healthcare costs. Many computational methods already predict interactions between small molecule drugs (SMDs). As the number of biotechnology drugs (BioDs) increases, so makes the threat of interactions between SMDs and BioDs. However, few computational methods are available to predict their interactions. RESULTS: Considering the structural specificity and relational complexity of SMDs and BioDs, a novel multi-modal representation learning method called Multi-SBI is proposed to predict their interactions. First, multi-modal features are used to adequately represent the heterogeneous structure and complex relationships of SMDs and BioDs. Second, an undersampling method based on Positive-unlabeled learning (PU-sampling) is introduced to obtain negative samples with high confidence from the unlabeled data set. Finally, both learned representations of SMD and BioD are fed into DNN classifiers to predict their interaction events. In addition, we also conduct a retrospective analysis. CONCLUSIONS: Our proposed multi-modal representation learning method can extract drug features more comprehensively in heterogeneous drugs. In addition, PU-sampling can effectively reduce the noise in the sampling procedure. Our proposed method significantly outperforms other state-of-the-art drug interaction prediction methods. In a retrospective analysis of DrugBank 5.1.0, 14 out of the 20 predictions with the highest confidence were validated in the latest version of DrugBank 5.1.8, demonstrating that Multi-SBI is a valuable tool for predicting new drug interactions through effectively extracting and learning heterogeneous drug features.

BACKGROUND: RNA secondary structure has broad impact on the fate of RNA metabolism. The reduced stability of secondary structures near the translation initiation site/start codon of the coding region promotes the efficiency of translation in both prokaryotic and eukaryotic species. However, the inaccuracy of in silico folding and the focus on the coding region limit our understanding of the global relationship between the whole mRNA structure and translation efficiency. Leveraging high-throughput RNA structure probing data in the transcriptome, we aim to systematically investigate the role of RNA structure in regulating translation efficiency. RESULTS: Here, we analyze the influences of hundreds of sequence and structural features on translation efficiency in the mouse embryonic stem cells (mESCs) and zebrafish developmental stages. Our findings reveal that overall in vivo RNA structure has a higher relative importance in predicting translation efficiency than in vitro RNA structure in both mESCs and zebrafish. Also, RNA structures in 3' untranslated region (UTR) have much stronger influence on translation efficiency compared to those in coding regions or 5' UTR. Furthermore, strong alternation between in vitro and in vivo structures in 3' UTR are detected in highly translated mRNAs in mESCs but not zebrafish. Instead, moderate alteration between in vitro and in vivo RNA structures in the 5' UTR and proximal coding regions are detected in highly translated mRNAs in zebrafish. CONCLUSIONS: Our results suggest the openness of the 3' UTR promotes the translation efficiency in both mice and zebrafish, with the in vivo structure in 3' UTR more important in mice than in zebrafish. This reveals a novel role of RNA secondary structure on translational regulation.

BACKGROUND: Anticancer peptide (ACP) inhibits and kills tumor cells. Research on ACP is of great significance for the development of new drugs, and the prediction of ACPs and non-ACPs is the new hotspot. RESULTS: We propose a new machine learning-based method named GCNCPR-ACPs (a Graph Convolutional Neural Network Method based on collapse pooling and residual network to predict the ACPs), which automatically and accurately predicts ACPs using residual graph convolution networks, differentiable graph pooling, and features extracted using peptide sequence information extraction. The GCNCPR-ACPs method can effectively capture different levels of node attributes for amino acid node representation learning, GCNCPR-ACPs uses node2vec and one-hot embedding methods to extract initial amino acid features for ACP prediction. CONCLUSIONS: Experimental results of ten-fold cross-validation and independent validation based on different metrics showed that GCNCPR-ACPs significantly outperformed state-of-the-art methods. Specifically, the evaluation indicators of Matthews Correlation Coefficient (MCC) and AUC of our predicator were 69.5% and 90%, respectively, which were 4.3% and 2% higher than those of the other predictors, respectively, in ten-fold cross-validation. And in the independent test, the scores of MCC and SP were 69.6% and 93.9%, respectively, which were 37.6% and 5.5% higher than those of the other predictors, respectively. The overall results showed that the GCNCPR-ACPs method proposed in the current paper can effectively predict ACPs.

BACKGROUND: In order to detect threats to public health and to be well-prepared for endemic and pandemic illness outbreaks, countries usually rely on event-based surveillance (EBS) and indicator-based surveillance systems. Event-based surveillance systems are key components of early warning systems and focus on fast capturing of data to detect threat signals through channels other than traditional surveillance. In this study, we develop Natural Language Processing tools that can be used within EBS systems. In particular, we focus on information extraction techniques that enable digital surveillance to monitor Internet data and social media. RESULTS: We created an annotated Spanish corpus from ProMED-mail health reports regarding disease outbreaks in Latin America. The corpus has been used to train algorithms for two information extraction tasks: named entity recognition and relation extraction. The algorithms, based on deep learning and rules, have been applied to recognize diseases, hosts, and geographical locations where a disease is occurring, among other entities and relations. In addition, an in-depth analysis of micro-average F1 metrics shows the suitability of our approaches for both tasks. CONCLUSIONS: The annotated corpus and algorithms presented could leverage the development of automated tools for extracting information from news and health reports written in Spanish. Moreover, this framework could be useful within EBS systems to support the early detection of Latin American disease outbreaks.

Over the last few years, dozens of healthcare surveys have shown a shortage of doctors and an alarming doctor-population ratio. With the motivation of assisting doctors and utilizing their time efficiently, automatic disease diagnosis using artificial intelligence is experiencing an ever-growing demand and popularity. Humans are known by the company they keep; similarly, symptoms also exhibit the association property, i.e., one symptom may strongly suggest another symptom's existence/non-existence, and their association provides crucial information about the suffering condition. The work investigates the role of symptom association in symptom investigation and disease diagnosis process. We propose and build a virtual assistant called Association guided Symptom Investigation and Diagnosis Assistant (A-SIDA) using hierarchical reinforcement learning. The proposed A-SIDDA converses with patients and extracts signs and symptoms as per patients' chief complaints and ongoing dialogue context. We infused association-based recommendations and critic into the assistant, which reinforces the assistant for conducting context-aware, symptom-association guided symptom investigation. Following the symptom investigation, the assistant diagnoses a disease based on the extracted signs and symptoms. The assistant then diagnoses a disease based on the extracted signs and symptoms. In addition to diagnosis accuracy, the relevance of inspected symptoms is critical to the usefulness of a diagnosis framework. We also propose a novel evaluation metric called Investigation Relevance Score (IReS), which measures the relevance of symptoms inspected during symptom investigation. The obtained improvements (Diagnosis success rate-5.36%, Dialogue length-1.16, Match rate-2.19%, Disease classifier-6.36%, IReS-0.3501, and Human score-0.66) over state-of-the-art methods firmly establish the crucial role of symptom association that gets uncovered by the virtual agent. Furthermore, we found that the association guided symptom investigation greatly increases human satisfaction, owing to its seamless topic (symptom) transition.

PURPOSE: Autism spectrum disorder (ASD) is the most prevalent disease today. The causes of its infection may be attributed to genetic causes by 80% and environmental causes by 20%. In spite of this, the majority of the current research is concerned with environmental causes, and the least proportion with the genetic causes of the disease. Autism is a complex disease, which makes it difficult to identify the genes that cause the disease. METHODS: Hybrid ensemble-based classification (HEC-ASD) model for predicting ASD genes using gradient boosting machines is proposed. The proposed model utilizes gene ontology (GO) to construct a gene functional similarity matrix using hybrid gene similarity (HGS) method. HGS measures the semantic similarity between genes effectively. It combines the graph-based method, such as Wang method with the number of directed children's nodes of gene term from GO. Moreover, an ensemble gradient boosting classifier is adapted to enhance the prediction of genes forming a robust classification model. RESULTS: The proposed model is evaluated using the Simons Foundation Autism Research Initiative (SFARI) gene database. The experimental results are promising as they improve the classification performance for predicting ASD genes. The results are compared with other approaches that used gene regulatory network (GRN), protein to protein interaction network (PPI), or GO. The HEC-ASD model reaches the highest prediction accuracy of 0.88% using ensemble learning classifiers. CONCLUSION: The proposed model demonstrates that ensemble learning technique using gradient boosting is effective in predicting autism spectrum disorder genes. Moreover, the HEC-ASD model utilized GO rather than using PPI network and GRN.

MOTIVATION: In eukaryotes, homologous recombination between the parental genomes frequently occurs during the evolutionary conserved process of meiosis, generating the genetic diversity transmitted by the gametes. The genome-wide determination of the frequency and location of the recombination events can now be efficiently performed by genotyping the offspring's polymorphic markers. However, genotyping recombination in complex hybrid genomes with existing methods remains challenging because of their strain and ploidy specificity and the degree of diversity and complexity of the parental genomes, especially in [Formula: see text] polyploids. RESULTS: We present UGDR, a pipeline to genotype the polymorphisms of complex hybrid yeast genomes. It is based on optimal mapping strategies of NGS reads, comparative analyses of the allelic ratio variation and read depth coverage. We tested the UGDR pipeline with sequencing reads from recombined hybrid diploid yeast strains and various clinical strains exhibiting different degrees of ploidy. UGDR allows to plot the markers distribution and recombination profile per chromosome. CONCLUSION: UGDR detects and plots recombination events in haploids and polyploid yeasts, which facilitates the discovery and understanding of the yeast genetic recombination map and identify new out-performing recombinants.

Flux balance analysis (FBA) is an optimization based approach to find the optimal steady state of a metabolic network, commonly of microorganisms such as yeast strains and Escherichia coli. However, the resulting solution from an FBA is typically not unique, as the optimization problem is, more often than not, degenerate. Flux variability analysis (FVA) is a method to determine the range of possible reaction fluxes that still satisfy, within some optimality factor, the original FBA problem. The resulting range of reaction fluxes can be utilized to determine metabolic reactions of high importance, amongst other analyses. In the literature, this has been done by solving [Formula: see text] linear programs (LPs), with n being the number of reactions in the metabolic network. However, FVA can be solved with less than [Formula: see text] LPs by utilizing the basic feasible solution property of bounded LPs to reduce the number of LPs that are needed to be solved. In this work, a new algorithm is proposed to solve FVA that requires less than [Formula: see text] LPs. The proposed algorithm is benchmarked on a problem set of 112 metabolic network models ranging from single cell organisms (iMM904, ect) to a human metabolic system (Recon3D). Showing a reduction in the number of LPs required to solve the FVA problem and thus the time to solve an FVA problem.

BACKGROUND: As a highly aggressive disease, cancer has been becoming the leading death cause around the world. Accurate prediction of the survival expectancy for cancer patients is significant, which can help clinicians make appropriate therapeutic schemes. With the high-throughput sequencing technology becoming more and more cost-effective, integrating multi-type genome-wide data has been a promising method in cancer survival prediction. Based on these genomic data, some data-integration methods for cancer survival prediction have been proposed. However, existing methods fail to simultaneously utilize feature information and structure information of multi-type genome-wide data. RESULTS: We propose a Multi-type Data Joint Learning (MDJL) approach based on multi-type genome-wide data, which comprehensively exploits feature information and structure information. Specifically, MDJL exploits correlation representations between any two data types by cross-correlation calculation for learning discriminant features. Moreover, based on the learned multiple correlation representations, MDJL constructs sample similarity matrices for capturing global and local structures across different data types. With the learned discriminant representation matrix and fused similarity matrix, MDJL constructs graph convolutional network with Cox loss for survival prediction. CONCLUSIONS: Experimental results demonstrate that our approach substantially outperforms established integrative methods and is effective for cancer survival prediction.

Extracting knowledge from heterogeneous data sources is fundamental for the construction of structured biomedical knowledge graphs (BKGs), where entities and relations are represented as nodes and edges in the graphs, respectively. Previous biomedical knowledge extraction methods simply considered limited entity types and relations by using a task-specific training set, which is insufficient for large-scale BKGs development and downstream task applications in different scenarios. To alleviate this issue, we propose a joint continual learning biomedical information extraction (JCBIE) network to extract entities and relations from different biomedical information datasets. By empirically studying different joint learning and continual learning strategies, the proposed JCBIE can learn and expand different types of entities and relations from different datasets. JCBIE uses two separated encoders in joint-feature extraction, hence can effectively avoid the feature confusion problem comparing with using one hard-parameter sharing encoder. Specifically, it allows us to adopt entity augmented inputs to establish the interaction between named entity recognition and relation extraction. Finally, a novel evaluation mechanism is proposed for measuring cross-corpus generalization errors, which was ignored by traditional evaluation methods. Our empirical studies show that JCBIE achieves promising performance when continual learning strategy is adopted with multiple corpora.

BACKGROUND: The genomes of SARS-CoV-2 are classified into variants, some of which are monitored as variants of concern (e.g. the Delta variant B.1.617.2 or Omicron variant B.1.1.529). Proportions of these variants circulating in a human population are typically estimated by large-scale sequencing of individual patient samples. Sequencing a mixture of SARS-CoV-2 RNA molecules from wastewater provides a cost-effective alternative, but requires methods for estimating variant proportions in a mixed sample. RESULTS: We propose a new method based on a probabilistic model of sequencing reads, capturing sequence diversity present within individual variants, as well as sequencing errors. The algorithm is implemented in an open source Python program called VirPool. We evaluate the accuracy of VirPool on several simulated and real sequencing data sets from both Illumina and nanopore sequencing platforms, including wastewater samples from Austria and France monitoring the onset of the Alpha variant. CONCLUSIONS: VirPool is a versatile tool for wastewater and other mixed-sample analysis that can handle both short- and long-read sequencing data. Our approach does not require pre-selection of characteristic mutations for variant profiles, it is able to use the entire length of reads instead of just the most informative positions, and can also capture haplotype dependencies within a single read.

BACKGROUND: Medication recommendation based on electronic medical record (EMR) is a research hot spot in smart healthcare. For developing computational medication recommendation methods based on EMR, an important challenge is the lack of a large number of longitudinal EMR data with time correlation. Faced with this challenge, this paper proposes a new EMR-based medication recommendation model called MR-KPA, which combines knowledge-enhanced pre-training with the deep adversarial network to improve medication recommendation from both feature representation and the fine-tuning process. Firstly, a knowledge-enhanced pre-training visit model is proposed to realize domain knowledge-based external feature fusion and pre-training-based internal feature mining for improving the feature representation. Secondly, a medication recommendation model based on the deep adversarial network is developed to optimize the fine-tuning process of pre-training visit model and alleviate over-fitting of model caused by the task gap between pre-training and recommendation. RESULT: The experimental results on EMRs from medical and health institutions in Hainan Province, China show that the proposed MR-KPA model can effectively improve the accuracy of medication recommendation on small-scale longitudinal EMR data compared with existing representative methods. CONCLUSION: The advantages of the proposed MR-KPA are mainly attributed to knowledge enhancement based on ontology embedding, the pre-training visit model and adversarial training. Each of these three optimizations is very effective for improving the capability of medication recommendation on small-scale longitudinal EMR data, and the pre-training visit model has the most significant improvement effect. These three optimizations are also complementary, and their integration makes the proposed MR-KPA model achieve the best recommendation effect.

BACKGROUND: Today's biomedical imaging technology has been able to present the morphological structure or functional metabolic information of organisms at different scale levels, such as organ, tissue, cell, molecule and gene. However, different imaging modes have different application scope, advantages and disadvantages. In order to improve the role of medical image in disease diagnosis, the fusion of biomedical image information at different imaging modes and scales has become an important research direction in medical image. Traditional medical image fusion methods are all designed to measure the activity level and fusion rules. They are lack of mining the context features of different modes of image, which leads to the obstruction of improving the quality of fused images. METHOD: In this paper, an attention-multiscale network medical image fusion model based on contextual features is proposed. The model selects five backbone modules in the VGG-16 network to build encoders to obtain the contextual features of medical images. It builds the attention mechanism branch to complete the fusion of global contextual features and designs the residual multiscale detail processing branch to complete the fusion of local contextual features. Finally, it completes the cascade reconstruction of features by the decoder to obtain the fused image. RESULTS: Ten sets of images related to five diseases are selected from the AANLIB database to validate the VANet model. Structural images are derived from MR images with high resolution and functional images are derived from SPECT and PET images that are good at describing organ blood flow levels and tissue metabolism. Fusion experiments are performed on twelve fusion algorithms including the VANet model. The model selects eight metrics from different aspects to build a fusion quality evaluation system to complete the performance evaluation of the fused images. Friedman's test and the post-hoc Nemenyi test are introduced to conduct professional statistical tests to demonstrate the superiority of VANet model. CONCLUSIONS: The VANet model completely captures and fuses the texture details and color information of the source images. From the fusion results, the metabolism and structural information of the model are well expressed and there is no interference of color information on the structure and texture; in terms of the objective evaluation system, the metric value of the VANet model is generally higher than that of other methods.; in terms of efficiency, the time consumption of the model is acceptable; in terms of scalability, the model is not affected by the input order of source images and can be extended to tri-modal fusion.

As of June 2022, the GISAID database contains more than 11 million SARS-CoV-2 genomes, including several thousand nucleotide sequences for the most common variants such as delta or omicron. These SARS-CoV-2 strains have been collected from patients around the world since the beginning of the pandemic. We start by assessing the similarity of all pairs of nucleotide sequences using the Jaccard index and principal component analysis. As shown previously in the literature, an unsupervised cluster analysis applied to the SARS-CoV-2 genomes results in clusters of sequences according to certain characteristics such as their strain or their clade. Importantly, we observe that nucleotide sequences of common variants are often outliers in clusters of sequences stemming from variants identified earlier on during the pandemic. Motivated by this finding, we are interested in applying outlier detection to nucleotide sequences. We demonstrate that nucleotide sequences of common variants (such as alpha, delta, or omicron) can be identified solely based on a statistical outlier criterion. We argue that outlier detection might be a useful surveillance tool to identify emerging variants in real time as the pandemic progresses.

BACKGROUND: The human protein atlas (HPA) is an online database containing large sets of protein expression data in normal and cancerous tissues in image form from immunohistochemically (IHC) stained tissue microarrays. In these, the tissue architecture is preserved and thus provides information on the spatial distribution and localization of protein expression at the cellular and extracellular levels. The database is freely available online through the HPA website but currently without support for large-scale screening and analysis of the images in the database. Features like spatial information are typically lacking in gene expression datasets from homogenized tissues or single-cell analysis. To enable high throughput analysis of the HPA database, we developed the AtlasGrabber software. It is available freely under an open-source license. Based on a predefined gene list, the software fetches the images from the database and displays them for the user. Several filters for specific antibodies or images enable the user to customize her/his image analysis. Up to four images can be displayed simultaneously, which allows for the comparison of protein expression between different tissues and between normal and cancerous tissues. An additional feature is the XML parser that allows the extraction of a list of available antibodies, images, and genes for specific tissues or cancer types from the HPA's database file. RESULTS: Compared to existing software designed for a similar purpose, ours provide more functionality and is easier to use. To demonstrate the software's usability, we identified six new markers of basal cells of the prostate. A comparison to prostate cancer showed that five of them are absent in prostate cancer. CONCLUSIONS: The HPA is a uniquely valuable database. By facilitating its usefulness with the AtlasGrabber, we enable researchers to exploit its full capacity. The loss of basal cell markers is diagnostic for prostate cancer and can help refine the histopathological diagnosis of prostate cancer. As proof of concept, with the AtlasGrabber we identified five new potential biomarkers specific for prostate basal cells which are lost in prostate cancer and thus can be used for prostate cancer diagnostics.

BACKGROUND: The Basic Local Alignment Search Tool (BLAST) is a suite of commonly used algorithms for identifying matches between biological sequences. The user supplies a database file and query file of sequences for BLAST to find identical sequences between the two. The typical millions of database and query sequences make BLAST computationally challenging but also well suited for parallelization on high-performance computing clusters. The efficacy of parallelization depends on the data partitioning, where the optimal data partitioning relies on an accurate performance model. In previous studies, a BLAST job was sped up by 27 times by partitioning the database and query among thousands of processor nodes. However, the optimality of the partitioning method was not studied. Unlike BLAST performance models proposed in the literature that usually have problem size and hardware configuration as the only variables, the execution time of a BLAST job is a function of database size, query size, and hardware capability. In this work, the nucleotide BLAST application BLASTN was profiled using three methods: shell-level profiling with the Unix "time" command, code-level profiling with the built-in "profiler" module, and system-level profiling with the Unix "gprof" program. The runtimes were measured for six node types, using six different database files and 15 query files, on a heterogeneous HPC cluster with 500+ nodes. The empirical measurement data were fitted with quadratic functions to develop performance models that were used to guide the data parallelization for BLASTN jobs. RESULTS: Profiling results showed that BLASTN contains more than 34,500 different functions, but a single function, RunMTBySplitDB, takes 99.12% of the total runtime. Among its 53 child functions, five core functions were identified to make up 92.12% of the overall BLASTN runtime. Based on the performance models, static load balancing algorithms can be applied to the BLASTN input data to minimize the runtime of the longest job on an HPC cluster. Four test cases being run on homogeneous and heterogeneous clusters were tested. Experiment results showed that the runtime can be reduced by 81% on a homogeneous cluster and by 20% on a heterogeneous cluster by re-distributing the workload. DISCUSSION: Optimal data partitioning can improve BLASTN's overall runtime 5.4-fold in comparison with dividing the database and query into the same number of fragments. The proposed methodology can be used in the other applications in the BLAST+ suite or any other application as long as source code is available.

BACKGROUND: Compound-protein interaction site and binding affinity predictions are crucial for drug discovery and drug design. In recent years, many deep learning-based methods have been proposed for predications related to compound-protein interaction. For protein inputs, how to make use of protein primary sequence and tertiary structure information has impact on prediction results. RESULTS: In this study, we propose a deep learning model based on a multi-objective neural network, which involves a multi-objective neural network for compound-protein interaction site and binding affinity prediction. We used several kinds of self-supervised protein embeddings to enrich our protein inputs and used convolutional neural networks to extract features from them. Our results demonstrate that our model had improvements in terms of interaction site prediction and affinity prediction compared to previous models. In a case study, our model could better predict binding sites, which also showed its effectiveness. CONCLUSION: These results suggest that our model could be a helpful tool for compound-protein related predictions.

BACKGROUND: Human subtelomeric DNA regulates the length and stability of adjacent telomeres that are critical for cellular function, and contains many gene/pseudogene families. Large evolutionarily recent segmental duplications and associated structural variation in human subtelomeres has made complete sequencing and assembly of these regions difficult to impossible for many loci, complicating or precluding a wide range of genetic analyses to investigate their function. RESULTS: We present a hybrid assembly method, NanoPore Guided REgional Assembly Tool (NPGREAT), which combines Linked-Read data with mapped ultralong nanopore reads spanning subtelomeric segmental duplications to potentially overcome these difficulties. Linked-Read sets of DNA sequences identified by matches with 1-copy subtelomere sequence adjacent to segmental duplications are assembled and extended into the segmental duplication regions using Regional Extension of Assemblies using Linked-Reads (REXTAL). Mapped telomere-containing ultralong nanopore reads are then used to provide contiguity and correct orientation for matching REXTAL sequence contigs as well as identification/correction of any misassemblies. Our method was tested for a subset of representative subtelomeres with ultralong nanopore read coverage in the haploid human cell line CHM13. A 10X Linked-Read dataset from CHM13 was combined with ultralong nanopore reads from the same genome to provide improved subtelomere assemblies. Comparison of Nanopore-only assemblies using SHASTA with our NPGREAT assemblies in the distal-most subtelomere regions showed that NPGREAT produced higher-quality and more complete assemblies than SHASTA alone when these regions had low ultralong nanopore coverage (such as cases where large segmental duplications were immediately adjacent to (TTAGGG) tracts). CONCLUSION: In genomic regions with large segmental duplications adjacent to telomeres, NPGREAT offers an alternative economical approach to improving assembly accuracy and coverage using linked-read datasets when more expensive HiFi datasets of 10-20 kb reads are unavailable.

During atherosclerosis, the narrowing of the arterial lumen is observed through the accumulation of bio compounds and the formation of plaque within artery walls. A non-linear optical imaging modality (NLOM), coherent anti-stokes Raman scattering (CARS) microscopy, can be used to image lipid-rich structures commonly found in atherosclerotic plaques. By matching the lipid's molecular vibrational frequencies (CH bonds), it is possible to map the accumulation of lipid-rich structures without the need for exogenous labelling and/or processing of the samples. CARS allows for the visualization of the morphological features of plaque. In combination with supervised machine learning, CARS imaged morphological features can be used to characterize the progression of atherosclerotic plaques.  RESULTS: Based on a set of label-free CARS images of atherosclerotic plaques (i.e. foam cell clusters) from a Watanabe heritable hyperlipidemic rabbit model, we developed an automated pipeline to classify atherosclerotic lesions based on their major morphological features. Our method uses image preprocessing to first improve the quality of the CARS-imaged plaque, followed by the segmentation of the plaque using Otsu thresholding, marker-controlled watershed, K-means segmentation and a novel independent foam cell thresholding segmentation. To define relevant morphological features, 27 quantitative features were extracted and further refined by a novel coefficient of variation feature refinement method in accordance with filter-type feature selection. Refined morphological features were supplied into three supervised machine learning algorithms; K-nearest neighbour, support vector machine and decision tree classifier. The classification pipeline showcased the ability to exploit relevant plaque morphological features to accurately classify 3 pre-defined stages of atherosclerosis: early fatty streak development (EFS) and advancing atheroma (AA) with a greater than 85% class accuracy CONCLUSIONS: Through the combination of CARS microscopy and computational methods, a powerful classification tool was developed to identify the progression of atherosclerotic plaque in an automated manner. Using a curated dataset, the classification pipeline demonstrated the ability to differentiate between EFS, EF and AA. Thus, presenting the opportunity to classify the onset of atherosclerosis at an earlier stage of development.

BACKGROUND: Brain tumor segmentation plays a significant role in clinical treatment and surgical planning. Recently, several deep convolutional networks have been proposed for brain tumor segmentation and have achieved impressive performance. However, most state-of-the-art models use 3D convolution networks, which require high computational costs. This makes it difficult to apply these models to medical equipment in the future. Additionally, due to the large diversity of the brain tumor and uncertain boundaries between sub-regions, some models cannot well-segment multiple tumors in the brain at the same time. RESULTS: In this paper, we proposed a lightweight hierarchical convolution network, called LHC-Net. Our network uses a multi-scale strategy which the common 3D convolution is replaced by the hierarchical convolution with residual-like connections. It improves the ability of multi-scale feature extraction and greatly reduces parameters and computation resources. On the BraTS2020 dataset, LHC-Net achieves the Dice scores of 76.38%, 90.01% and 83.32% for ET, WT and TC, respectively, which is better than that of 3D U-Net with 73.50%, 89.42% and 81.92%. Especially on the multi-tumor set, our model shows significant performance improvement. In addition, LHC-Net has 1.65M parameters and 35.58G FLOPs, which is two times fewer parameters and three times less computation compared with 3D U-Net. CONCLUSION: Our proposed method achieves automatic segmentation of tumor sub-regions from four-modal brain MRI images. LHC-Net achieves competitive segmentation performance with fewer parameters and less computation than the state-of-the-art models. It means that our model can be applied under limited medical computing resources. By using the multi-scale strategy on channels, LHC-Net can well-segment multiple tumors in the patient's brain. It has great potential for application to other multi-scale segmentation tasks.

BACKGROUND: Long-read shotgun metagenomic sequencing is gaining in popularity and offers many advantages over short-read sequencing. The higher information content in long reads is useful for a variety of metagenomics analyses, including taxonomic classification and profiling. The development of long-read specific tools for taxonomic classification is accelerating, yet there is a lack of information regarding their relative performance. Here, we perform a critical benchmarking study using 11 methods, including five methods designed specifically for long reads. We applied these tools to several mock community datasets generated using Pacific Biosciences (PacBio) HiFi or Oxford Nanopore Technology sequencing, and evaluated their performance based on read utilization, detection metrics, and relative abundance estimates. RESULTS: Our results show that long-read classifiers generally performed best. Several short-read classification and profiling methods produced many false positives (particularly at lower abundances), required heavy filtering to achieve acceptable precision (at the cost of reduced recall), and produced inaccurate abundance estimates. By contrast, two long-read methods (BugSeq, MEGAN-LR & DIAMOND) and one generalized method (sourmash) displayed high precision and recall without any filtering required. Furthermore, in the PacBio HiFi datasets these methods detected all species down to the 0.1% abundance level with high precision. Some long-read methods, such as MetaMaps and MMseqs2, required moderate filtering to reduce false positives to resemble the precision and recall of the top-performing methods. We found read quality affected performance for methods relying on protein prediction or exact k-mer matching, and these methods performed better with PacBio HiFi datasets. We also found that long-read datasets with a large proportion of shorter reads (< 2 kb length) resulted in lower precision and worse abundance estimates, relative to length-filtered datasets. Finally, for classification methods, we found that the long-read datasets produced significantly better results than short-read datasets, demonstrating clear advantages for long-read metagenomic sequencing. CONCLUSIONS: Our critical assessment of available methods provides best-practice recommendations for current research using long reads and establishes a baseline for future benchmarking studies.

BACKGROUND: Automatic functional annotation of proteins is an open research problem in bioinformatics. The growing number of protein entries in public databases, for example in UniProtKB, poses challenges in manual functional annotation. Manual annotation requires expert human curators to search and read related research articles, interpret the results, and assign the annotations to the proteins. Thus, it is a time-consuming and expensive process. Therefore, designing computational tools to perform automatic annotation leveraging the high quality manual annotations that already exist in UniProtKB/SwissProt is an important research problem RESULTS: In this paper, we extend and adapt the GrAPFI (graph-based automatic protein function inference) (Sarker et al. in BMC Bioinform 21, 2020; Sarker et al., in: Proceedings of 7th international conference on complex networks and their applications, Cambridge, 2018) method for automatic annotation of proteins with gene ontology (GO) terms renaming it as GrAPFI-GO. The original GrAPFI method uses label propagation in a similarity graph where proteins are linked through the domains, families, and superfamilies that they share. Here, we also explore various types of similarity measures based on common neighbors in the graph. Moreover, GO terms are arranged in a hierarchical manner according to semantic parent-child relations. Therefore, we propose an efficient pruning and post-processing technique that integrates both semantic similarity and hierarchical relations between the GO terms. We produce experimental results comparing the GrAPFI-GO method with and without considering common neighbors similarity. We also test the performance of GrAPFI-GO and other annotation tools for GO annotation on a benchmark of proteins with and without the proposed pruning and post-processing procedure. CONCLUSION: Our results show that the proposed semantic hierarchical post-processing potentially improves the performance of GrAPFI-GO and of other annotation tools as well. Thus, GrAPFI-GO exposes an original efficient and reusable procedure, to exploit the semantic relations among the GO terms in order to improve the automatic annotation of protein functions.

BACKGROUND: Today it is possible to profile the transcriptome of individual cells, and a key step in the analysis of these datasets is unsupervised clustering. For very large datasets, efficient algorithms are required to ensure that analyses can be conducted with reasonable time and memory requirements. RESULTS: Here, we present a highly efficient k-means based approach, and we demonstrate that it scales favorably with the number of cells with regards to time and memory. CONCLUSIONS: We have demonstrated that our streaming k-means clustering algorithm gives state-of-the-art performance while resource requirements scale favorably for up to 2 million cells.

BACKGROUND: Differential abundance testing is an important aspect of microbiome data analysis, where each taxa is fitted with a statistical test or a regression model. However, many models do not provide a good fit to real microbiome data. This has been shown to result in high false positive rates. Permutation tests are a good alternative, but a regression approach is desired for small data sets with many covariates, where stratification is not an option. RESULTS: We implement an R package 'llperm' where the The Permutation of Regressor Residuals (PRR) test can be applied to any likelihood based model, not only generalized linear models. This enables distributions with zero-inflation and overdispersion, making the test suitable for count regression models popular in microbiome data analysis. Simulations based on a real data set show that the PRR-test approach is able to maintain the correct nominal false positive rate expected from the null hypothesis, while having equal or greater power to detect the true positives as models based on likelihood at a given false positive rate. CONCLUSIONS: Standard count regression models can have a shockingly high false positive rate in microbiome data sets. As they may lead to false conclusions, the guaranteed nominal false positive rate gained from the PRR-test can be viewed as a major benefit.

BACKGROUND: Investigating molecular heterogeneity provides insights into tumour origin and metabolomics. The increasing amount of data gathered makes manual analyses infeasible-therefore, automated unsupervised learning approaches are utilised for discovering tissue heterogeneity. However, automated analyses require experience setting the algorithms' hyperparameters and expert knowledge about the analysed biological processes. Moreover, feature engineering is needed to obtain valuable results because of the numerous features measured. RESULTS: We propose DiviK: a scalable stepwise algorithm with local data-driven feature space adaptation for segmenting high-dimensional datasets. The algorithm is compared to the optional solutions (regular k-means, spatial and spectral approaches) combined with different feature engineering techniques (None, PCA, EXIMS, UMAP, Neural Ions). Three quality indices: Dice Index, Rand Index and EXIMS score, focusing on the overall composition of the clustering, coverage of the tumour region and spatial cluster consistency, are used to assess the quality of unsupervised analyses. Algorithms were validated on mass spectrometry imaging (MSI) datasets-2D human cancer tissue samples and 3D mouse kidney images. DiviK algorithm performed the best among the four clustering algorithms compared (overall quality score 1.24, 0.58 and 162 for d(0, 0, 0), d(1, 1, 1) and the sum of ranks, respectively), with spectral clustering being mostly second. Feature engineering techniques impact the overall clustering results less than the algorithms themselves (partial [Formula: see text] effect size: 0.141 versus 0.345, Kendall's concordance index: 0.424 versus 0.138 for d(0, 0, 0)). CONCLUSIONS: DiviK could be the default choice in the exploration of MSI data. Thanks to its unique, GMM-based local optimisation of the feature space and deglomerative schema, DiviK results do not strongly depend on the feature engineering technique applied and can reveal the hidden structure in a tissue sample. Additionally, DiviK shows high scalability, and it can process at once the big omics data with more than 1.5 mln instances and a few thousand features. Finally, due to its simplicity, DiviK is easily generalisable to an even more flexible framework. Therefore, it is helpful for other -omics data (as single cell spatial transcriptomic) or tabular data in general (including medical images after appropriate embedding). A generic implementation is freely available under Apache 2.0 license at https://github.com/gmrukwa/divik .

BACKGROUND: The widespread usage of Cap Analysis of Gene Expression (CAGE) has led to numerous breakthroughs in understanding the transcription mechanisms. Recent evidence in the literature, however, suggests that CAGE suffers from transcriptional and technical noise. Regardless of the sample quality, there is a significant number of CAGE peaks that are not associated with transcription initiation events. This type of signal is typically attributed to technical noise and more frequently to random five-prime capping or transcription bioproducts. Thus, the need for computational methods emerges, that can accurately increase the signal-to-noise ratio in CAGE data, resulting in error-free transcription start site (TSS) annotation and quantification of regulatory region usage. In this study, we present DeepTSS, a novel computational method for processing CAGE samples, that combines genomic signal processing (GSP), structural DNA features, evolutionary conservation evidence and raw DNA sequence with Deep Learning (DL) to provide single-nucleotide TSS predictions with unprecedented levels of performance. RESULTS: To evaluate DeepTSS, we utilized experimental data, protein-coding gene annotations and computationally-derived genome segmentations by chromatin states. DeepTSS was found to outperform existing algorithms on all benchmarks, achieving 98% precision and 96% sensitivity (accuracy 95.4%) on the protein-coding gene strategy, with 96.66% of its positive predictions overlapping active chromatin, 98.27% and 92.04% co-localized with at least one transcription factor and H3K4me3 peak. CONCLUSIONS: CAGE is a key protocol in deciphering the language of transcription, however, as every experimental protocol, it suffers from biological and technical noise that can severely affect downstream analyses. DeepTSS is a novel DL-based method for effectively removing noisy CAGE signal. In contrast to existing software, DeepTSS does not require feature selection since the embedded convolutional layers can readily identify patterns and only utilize the important ones for the classification task. This study highlights the key role that DL can play in Molecular Biology, by removing the inherent flaws of experimental protocols, that form the backbone of contemporary research. Here, we show how DeepTSS can unleash the full potential of an already popular and mature method such as CAGE, and push the boundaries of coding and non-coding gene expression regulator research even further.

BACKGROUND: Medical databases normally contain large amounts of data in a variety of forms. Although they grant significant insights into diagnosis and treatment, implementing data exploration into current medical databases is challenging since these are often based on a relational schema and cannot be used to easily extract information for cohort analysis and visualization. As a consequence, valuable information regarding cohort distribution or patient similarity may be missed. With the rapid advancement of biomedical technologies, new forms of data from methods such as Next Generation Sequencing (NGS) or chromosome microarray (array CGH) are constantly being generated; hence it can be expected that the amount and complexity of medical data will rise and bring relational database systems to a limit. DESCRIPTION: We present Graph4Med, a web application that relies on a graph database obtained by transforming a relational database. Graph4Med provides a straightforward visualization and analysis of a selected patient cohort. Our use case is a database of pediatric Acute Lymphoblastic Leukemia (ALL). Along routine patients' health records it also contains results of latest technologies such as NGS data. We developed a suitable graph data schema to convert the relational data into a graph data structure and store it in Neo4j. We used NeoDash to build a dashboard for querying and displaying patients' cohort analysis. This way our tool (1) quickly displays the overview of patients' cohort information such as distributions of gender, age, mutations (fusions), diagnosis; (2) provides mutation (fusion) based similarity search and display in a maneuverable graph; (3) generates an interactive graph of any selected patient and facilitates the identification of interesting patterns among patients. CONCLUSION: We demonstrate the feasibility and advantages of a graph database for storing and querying medical databases. Our dashboard allows a fast and interactive analysis and visualization of complex medical data. It is especially useful for patients similarity search based on mutations (fusions), of which vast amounts of data have been generated by NGS in recent years. It can discover relationships and patterns in patients cohorts that are normally hard to grasp. Expanding Graph4Med to more medical databases will bring novel insights into diagnostic and research.

BACKGROUND: Cis-regulatory regions (CRRs) are non-coding regions of the DNA that fine control the spatio-temporal pattern of transcription; they are involved in a wide range of pivotal processes such as the development of specific cell-lines/tissues and the dynamic cell response to physiological stimuli. Recent studies showed that genetic variants occurring in CRRs are strongly correlated with pathogenicity or deleteriousness. Considering the central role of CRRs in the regulation of physiological and pathological conditions, the correct identification of CRRs and of their tissue-specific activity status through Machine Learning methods plays a major role in dissecting the impact of genetic variants on human diseases. Unfortunately, the problem is still open, though some promising results have been already reported by (deep) machine-learning based methods that predict active promoters and enhancers in specific tissues or cell lines by encoding epigenetic or spectral features directly extracted from DNA sequences. RESULTS: We present the experiments we performed to compare two Deep Neural Networks, a Feed-Forward Neural Network model working on epigenomic features, and a Convolutional Neural Network model working only on genomic sequence, targeted to the identification of enhancer- and promoter-activity in specific cell lines. While performing experiments to understand how the experimental setup influences the prediction performance of the methods, we particularly focused on (1) automatic model selection performed by Bayesian optimization and (2) exploring different data rebalancing setups for reducing negative unbalancing effects. CONCLUSIONS: Results show that (1) automatic model selection by Bayesian optimization improves the quality of the learner; (2) data rebalancing considerably impacts the prediction performance of the models; test set rebalancing may provide over-optimistic results, and should therefore be cautiously applied; (3) despite working on sequence data, convolutional models obtain performance close to those of feed forward models working on epigenomic information, which suggests that also sequence data carries informative content for CRR-activity prediction. We therefore suggest combining both models/data types in future works.

BACKGROUND: With the rapid accumulation of scRNA-seq data, more and more automatic cell type identification methods have been developed, especially those based on deep learning. Although these methods have reached relatively high prediction accuracy, many issues still exist. One is the interpretability. The second is how to deal with the non-standard test samples that are not encountered in the training process. RESULTS: Here we introduce scCapsNet-mask, an updated version of scCapsNet. The scCapsNet-mask provides a reasonable solution to the issues of interpretability and non-standard test samples. Firstly, the scCapsNet-mask utilizes a mask to ease the task of model interpretation in the original scCapsNet. The results show that scCapsNet-mask could constrain the coupling coefficients, and make a one-to-one correspondence between the primary capsules and type capsules. Secondly, the scCapsNet-mask can process non-standard samples more reasonably. In one example, the scCapsNet-mask was trained on the committed cells, and then tested on less differentiated cells as the non-standard samples. It could not only estimate the lineage bias of less differentiated cells, but also distinguish the development stages more accurately than traditional machine learning models. Therefore, the pseudo-temporal order of cells for each lineage could be established. Following these pseudo-temporal order, lineage specific genes exhibit a gradual increase expression pattern and stem cell associated genes exhibit a gradual decrease expression pattern. In another example, the scCapsNet-mask was trained on scRNA-seq data, and then used to assign cell type in spatial transcriptomics that may contain non-standard sample of doublets. The results show that the scCapsNet-mask not only restored the spatial map but also identified several non-standard samples of doublet. CONCLUSIONS: The scCapsNet-mask offers a suitable solution to the challenge of interpretability and non-standard test samples. By adding a mask, it has the advantages of automatic processing and easy interpretation compared with the original scCapsNet. In addition, the scCapsNet-mask could more accurately reflect the composition of non-standard test samples than traditional machine learning methods. Therefore, it can extend its applicability in functional analysis, such as fate bias prediction in less differentiated cells and cell type assignment in spatial transcriptomics.

BACKGROUND: This research aims to increase our knowledge of amyloidoses. These disorders cause incorrect protein folding, affecting protein functionality (on structure). Fibrillar deposits are the basis of some wellknown diseases, such as Alzheimer, Creutzfeldt-Jakob diseases and type II diabetes. For many of these amyloid proteins, the relative precursors are known. Discovering new protein precursors involved in forming amyloid fibril deposits would improve understanding the pathological processes of amyloidoses. RESULTS: A new classifier, called ENTAIL, was developed using over than 4000 molecular descriptors. ENTAIL was based on the Naive Bayes Classifier with Unbounded Support and Gaussian Kernel Type, with an accuracy on the test set of 81.80%, SN of 100%, SP of 63.63% and an MCC of 0.683 on a balanced dataset. CONCLUSIONS: The analysis carried out has demonstrated how, despite the various configurations of the tests, performances are superior in terms of performance on a balanced dataset.

BACKGROUND: Drug repositioning is a very important task that provides critical information for exploring the potential efficacy of drugs. Yet developing computational models that can effectively predict drug-disease associations (DDAs) is still a challenging task. Previous studies suggest that the accuracy of DDA prediction can be improved by integrating different types of biological features. But how to conduct an effective integration remains a challenging problem for accurately discovering new indications for approved drugs. METHODS: In this paper, we propose a novel meta-path based graph representation learning model, namely RLFDDA, to predict potential DDAs on heterogeneous biological networks. RLFDDA first calculates drug-drug similarities and disease-disease similarities as the intrinsic biological features of drugs and diseases. A heterogeneous network is then constructed by integrating DDAs, disease-protein associations and drug-protein associations. With such a network, RLFDDA adopts a meta-path random walk model to learn the latent representations of drugs and diseases, which are concatenated to construct joint representations of drug-disease associations. As the last step, we employ the random forest classifier to predict potential DDAs with their joint representations. RESULTS: To demonstrate the effectiveness of RLFDDA, we have conducted a series of experiments on two benchmark datasets by following a ten-fold cross-validation scheme. The results show that RLFDDA yields the best performance in terms of AUC and F1-score when compared with several state-of-the-art DDAs prediction models. We have also conducted a case study on two common diseases, i.e., paclitaxel and lung tumors, and found that 7 out of top-10 diseases and 8 out of top-10 drugs have already been validated for paclitaxel and lung tumors respectively with literature evidence. Hence, the promising performance of RLFDDA may provide a new perspective for novel DDAs discovery over heterogeneous networks.

BACKGROUND: Self-interacting proteins (SIPs), two or more copies of the protein that can interact with each other expressed by one gene, play a central role in the regulation of most living cells and cellular functions. Although numerous SIPs data can be provided by using high-throughput experimental techniques, there are still several shortcomings such as in time-consuming, costly, inefficient, and inherently high in false-positive rates, for the experimental identification of SIPs even nowadays. Therefore, it is more and more significant how to develop efficient and accurate automatic approaches as a supplement of experimental methods for assisting and accelerating the study of predicting SIPs from protein sequence information. RESULTS: In this paper, we present a novel framework, termed GLCM-WSRC (gray level co-occurrence matrix-weighted sparse representation based classification), for predicting SIPs automatically based on protein evolutionary information from protein primary sequences. More specifically, we firstly convert the protein sequence into Position Specific Scoring Matrix (PSSM) containing protein sequence evolutionary information, exploiting the Position Specific Iterated BLAST (PSI-BLAST) tool. Secondly, using an efficient feature extraction approach, i.e., GLCM, we extract abstract salient and invariant feature vectors from the PSSM, and then perform a pre-processing operation, the adaptive synthetic (ADASYN) technique, to balance the SIPs dataset to generate new feature vectors for classification. Finally, we employ an efficient and reliable WSRC model to identify SIPs according to the known information of self-interacting and non-interacting proteins. CONCLUSIONS: Extensive experimental results show that the proposed approach exhibits high prediction performance with 98.10% accuracy on the yeast dataset, and 91.51% accuracy on the human dataset, which further reveals that the proposed model could be a useful tool for large-scale self-interacting protein prediction and other bioinformatics tasks detection in the future.