BACKGROUND: The accurate prediction of biological features from genomic data is paramount for precision medicine and sustainable agriculture. For decades, neural network models have been widely popular in fields like computer vision, astrophysics and targeted marketing given their prediction accuracy and their robust performance under big data settings. Yet neural network models have not made a successful transition into the medical and biological world due to the ubiquitous characteristics of biological data such as modest sample sizes, sparsity, and extreme heterogeneity. RESULTS: Here, we investigate the robustness, generalization potential and prediction accuracy of widely used convolutional neural network and natural language processing models with a variety of heterogeneous genomic datasets. Mainly, recurrent neural network models outperform convolutional neural network models in terms of prediction accuracy, overfitting and transferability across the datasets under study. CONCLUSIONS: While the perspective of a robust out-of-the-box neural network model is out of reach, we identify certain model characteristics that translate well across datasets and could serve as a baseline model for translational researchers.

OBJECTIVES: Immune microenvironment was closely related to the occurrence and progression of colorectal cancer (CRC). The objective of the current research was to develop and verify a Machine learning survival predictive system for CRC based on immune gene expression data and machine learning algorithms. METHODS: The current study performed differentially expressed analyses between normal tissues and tumor tissues. Univariate Cox regression was used to screen prognostic markers for CRC. Prognostic immune genes and transcription factors were used to construct an immune-related regulatory network. Three machine learning algorithms were used to create an Machine learning survival predictive system for CRC. Concordance indexes, calibration curves, and Brier scores were used to evaluate the performance of prognostic model. RESULTS: Twenty immune genes (BCL2L12, FKBP10, XKRX, WFS1, TESC, CCR7, SPACA3, LY6G6C, L1CAM, OSM, EXTL1, LY6D, FCRL5, MYEOV, FOXD1, REG3G, HAPLN1, MAOB, TNFSF11, and AMIGO3) were recognized as independent risk factors for CRC. A prognostic nomogram was developed based on the previous immune genes. Concordance indexes were 0.852, 0.778, and 0.818 for 1-, 3- and 5-year survival. This prognostic model could discriminate high risk patients with poor prognosis from low risk patients with favorable prognosis. CONCLUSIONS: The current study identified twenty prognostic immune genes for CRC patients and constructed an immune-related regulatory network. Based on three machine learning algorithms, the current research provided three individual mortality predictive curves. The Machine learning survival predictive system was available at: https://zhangzhiqiao8.shinyapps.io/Artificial_Intelligence_Survival_Prediction_for_CRC_B1005_1/ , which was valuable for individualized treatment decision before surgery.

BACKGROUND: The assembly task is an indispensable step in sequencing genomes of new organisms and studying structural genomic changes. In recent years, the dynamic development of next-generation sequencing (NGS) methods raises hopes for making whole-genome sequencing a fast and reliable tool used, for example, in medical diagnostics. However, this is hampered by the slowness and computational requirements of the current processing algorithms, which raises the need to develop more efficient algorithms. One possible approach, still little explored, is the use of quantum computing. RESULTS: We present a proof of concept of de novo assembly algorithm, using the Genomic Signal Processing approach, detecting overlaps between DNA reads by calculating the Pearson correlation coefficient and formulating the assembly problem as an optimization task (Traveling Salesman Problem). Computations performed on a classic computer were compared with the results achieved by a hybrid method combining CPU and QPU calculations. For this purpose quantum annealer by D-Wave was used. The experiments were performed with artificially generated data and DNA reads coming from a simulator, with actual organism genomes used as input sequences. To our knowledge, this work is one of the few where actual sequences of organisms were used to study the de novo assembly task on quantum annealer. CONCLUSIONS: Proof of concept carried out by us showed that the use of quantum annealer (QA) for the de novo assembly task might be a promising alternative to the computations performed in the classical model. The current computing power of the available devices requires a hybrid approach (combining CPU and QPU computations). The next step may be developing a hybrid algorithm strictly dedicated to the de novo assembly task, using its specificity (e.g. the sparsity and bounded degree of the overlap-layout-consensus graph).

BACKGROUND: Heterogeneous omics data, increasingly collected through high-throughput technologies, can contain hidden answers to very important and still unsolved biomedical questions. Their integration and processing are crucial mostly for tertiary analysis of Next Generation Sequencing data, although suitable big data strategies still address mainly primary and secondary analysis. Hence, there is a pressing need for algorithms specifically designed to explore big omics datasets, capable of ensuring scalability and interoperability, possibly relying on high-performance computing infrastructures. RESULTS: We propose RGMQL, a R/Bioconductor package conceived to provide a set of specialized functions to extract, combine, process and compare omics datasets and their metadata from different and differently localized sources. RGMQL is built over the GenoMetric Query Language (GMQL) data management and computational engine, and can leverage its open curated repository as well as its cloud-based resources, with the possibility of outsourcing computational tasks to GMQL remote services. Furthermore, it overcomes the limits of the GMQL declarative syntax, by guaranteeing a procedural approach in dealing with omics data within the R/Bioconductor environment. But mostly, it provides full interoperability with other packages of the R/Bioconductor framework and extensibility over the most used genomic data structures and processing functions. CONCLUSIONS: RGMQL is able to combine the query expressiveness and computational efficiency of GMQL with a complete processing flow in the R environment, being a fully integrated extension of the R/Bioconductor framework. Here we provide three fully reproducible example use cases of biological relevance that are particularly explanatory of its flexibility of use and interoperability with other R/Bioconductor packages. They show how RGMQL can easily scale up from local to parallel and cloud computing while it combines and analyzes heterogeneous omics data from local or remote datasets, both public and private, in a completely transparent way to the user.

BACKGROUND: As many interactions between the chemical and genomic space remain undiscovered, computational methods able to identify potential drug-target interactions (DTIs) are employed to accelerate drug discovery and reduce the required cost. Predicting new DTIs can leverage drug repurposing by identifying new targets for approved drugs. However, developing an accurate computational framework that can efficiently incorporate chemical and genomic spaces remains extremely demanding. A key issue is that most DTI predictions suffer from the lack of experimentally validated negative interactions or limited availability of target 3D structures. RESULTS: We report DT2Vec, a pipeline for DTI prediction based on graph embedding and gradient boosted tree classification. It maps drug-drug and protein-protein similarity networks to low-dimensional features and the DTI prediction is formulated as binary classification based on a strategy of concatenating the drug and target embedding vectors as input features. DT2Vec was compared with three top-performing graph similarity-based algorithms on a standard benchmark dataset and achieved competitive results. In order to explore credible novel DTIs, the model was applied to data from the ChEMBL repository that contain experimentally validated positive and negative interactions which yield a strong predictive model. Then, the developed model was applied to all possible unknown DTIs to predict new interactions. The applicability of DT2Vec as an effective method for drug repurposing is discussed through case studies and evaluation of some novel DTI predictions is undertaken using molecular docking. CONCLUSIONS: The proposed method was able to integrate and map chemical and genomic space into low-dimensional dense vectors and showed promising results in predicting novel DTIs.

BACKGROUND: Recently, automatically extracting biomedical relations has been a significant subject in biomedical research due to the rapid growth of biomedical literature. Since the adaptation to the biomedical domain, the transformer-based BERT models have produced leading results on many biomedical natural language processing tasks. In this work, we will explore the approaches to improve the BERT model for relation extraction tasks in both the pre-training and fine-tuning stages of its applications. In the pre-training stage, we add another level of BERT adaptation on sub-domain data to bridge the gap between domain knowledge and task-specific knowledge. Also, we propose methods to incorporate the ignored knowledge in the last layer of BERT to improve its fine-tuning. RESULTS: The experiment results demonstrate that our approaches for pre-training and fine-tuning can improve the BERT model performance. After combining the two proposed techniques, our approach outperforms the original BERT models with averaged F1 score improvement of 2.1% on relation extraction tasks. Moreover, our approach achieves state-of-the-art performance on three relation extraction benchmark datasets. CONCLUSIONS: The extra pre-training step on sub-domain data can help the BERT model generalization on specific tasks, and our proposed fine-tuning mechanism could utilize the knowledge in the last layer of BERT to boost the model performance. Furthermore, the combination of these two approaches further improves the performance of BERT model on the relation extraction tasks.

BACKGROUND: cfMeDIP-seq is a low-cost method for determining the DNA methylation status of cell-free DNA and it has been successfully combined with statistical methods for accurate cancer diagnostics. We investigate the diagnostic classification aspect by applying statistical tests and dimension reduction techniques for feature selection and probabilistic modeling for the cancer type classification, and we also study the effect of sequencing depth. METHODS: We experiment with a variety of statistical methods that use different feature selection and feature extraction methods as well as probabilistic classifiers for diagnostic decision making. We test the (moderated) t-tests and the Fisher's exact test for feature selection, principal component analysis (PCA) as well as iterative supervised PCA (ISPCA) for feature generation, and GLMnet and logistic regression methods with sparsity promoting priors for classification. Probabilistic programming language Stan is used to implement Bayesian inference for the probabilistic models. RESULTS AND CONCLUSIONS: We compare overlaps of differentially methylated genomic regions as chosen by different feature selection methods, and evaluate probabilistic classifiers by evaluating the area under the receiver operating characteristic scores on discovery and validation cohorts. While we observe that many methods perform equally well as, and occasionally considerably better than, GLMnet that was originally proposed for cfMeDIP-seq based cancer classification, we also observed that performance of different methods vary across sequencing depths, cancer types and study cohorts. Overall, methods that seem robust and promising include Fisher's exact test and ISPCA for feature selection as well as a simple logistic regression model with the number of hyper and hypo-methylated regions as features.

BACKGROUND: T cell receptors (TCRs) play critical roles in adaptive immune responses, and recent advances in genome technology have made it possible to examine the T cell receptor (TCR) repertoire at the individual sequence level. The analysis of the TCR repertoire with respect to clinical phenotypes can yield novel insights into the etiology and progression of immune-mediated diseases. However, methods for association analysis of the TCR repertoire have not been well developed. METHODS: We introduce an analysis tool, TCR-L, for evaluating the association between the TCR repertoire and disease outcomes. Our approach is developed under a mixed effect modeling, where the fixed effect represents features that can be explicitly extracted from TCR sequences while the random effect represents features that are hidden in TCR sequences and are difficult to be extracted. Statistical tests are developed to examine the two types of effects independently, and then the p values are combined. RESULTS: Simulation studies demonstrate that (1) the proposed approach can control the type I error well; and (2) the power of the proposed approach is greater than approaches that consider fixed effect only or random effect only. The analysis of real data from a skin cutaneous melanoma study identifies an association between the TCR repertoire and the short/long-term survival of patients. CONCLUSION: The TCR-L can accommodate features that can be extracted as well as features that are hidden in TCR sequences. TCR-L provides a powerful approach for identifying association between TCR repertoire and disease outcomes.

BACKGROUND: As many complex omics data have been generated during the last two decades, dimensionality reduction problem has been a challenging issue in better mining such data. The omics data typically consists of many features. Accordingly, many feature selection algorithms have been developed. The performance of those feature selection methods often varies by specific data, making the discovery and interpretation of results challenging. METHODS AND RESULTS: In this study, we performed a comprehensive comparative study of five widely used supervised feature selection methods (mRMR, INMIFS, DFS, SVM-RFE-CBR and VWMRmR) for multi-omics datasets. Specifically, we used five representative datasets: gene expression (Exp), exon expression (ExpExon), DNA methylation (hMethyl27), copy number variation (Gistic2), and pathway activity dataset (Paradigm IPLs) from a multi-omics study of acute myeloid leukemia (LAML) from The Cancer Genome Atlas (TCGA). The different feature subsets selected by the aforesaid five different feature selection algorithms are assessed using three evaluation criteria: (1) classification accuracy (Acc), (2) representation entropy (RE) and (3) redundancy rate (RR). Four different classifiers, viz., C4.5, NaiveBayes, KNN, and AdaBoost, were used to measure the classification accuary (Acc) for each selected feature subset. The VWMRmR algorithm obtains the best Acc for three datasets (ExpExon, hMethyl27 and Paradigm IPLs). The VWMRmR algorithm offers the best RR (obtained using normalized mutual information) for three datasets (Exp, Gistic2 and Paradigm IPLs), while it gives the best RR (obtained using Pearson correlation coefficient) for two datasets (Gistic2 and Paradigm IPLs). It also obtains the best RE for three datasets (Exp, Gistic2 and Paradigm IPLs). Overall, the VWMRmR algorithm yields best performance for all three evaluation criteria for majority of the datasets. In addition, we identified signature genes using supervised learning collected from the overlapped top feature set among five feature selection methods. We obtained a 7-gene signature (ZMIZ1, ENG, FGFR1, PAWR, KRT17, MPO and LAT2) for EXP, a 9-gene signature for ExpExon, a 7-gene signature for hMethyl27, one single-gene signature (PIK3CG) for Gistic2 and a 3-gene signature for Paradigm IPLs. CONCLUSION: We performed a comprehensive comparison of the performance evaluation of five well-known feature selection methods for mining features from various high-dimensional datasets. We identified signature genes using supervised learning for the specific omic data for the disease. The study will help incorporate higher order dependencies among features.

BACKGROUND: Histone Mark Modifications (HMs) are crucial actors in gene regulation, as they actively remodel chromatin to modulate transcriptional activity: aberrant combinatorial patterns of HMs have been connected with several diseases, including cancer. HMs are, however, reversible modifications: understanding their role in disease would allow the design of 'epigenetic drugs' for specific, non-invasive treatments. Standard statistical techniques were not entirely successful in extracting representative features from raw HM signals over gene locations. On the other hand, deep learning approaches allow for effective automatic feature extraction, but at the expense of model interpretation. RESULTS: Here, we propose ShallowChrome, a novel computational pipeline to model transcriptional regulation via HMs in both an accurate and interpretable way. We attain state-of-the-art results on the binary classification of gene transcriptional states over 56 cell-types from the REMC database, largely outperforming recent deep learning approaches. We interpret our models by extracting insightful gene-specific regulative patterns, and we analyse them for the specific case of the PAX5 gene over three differentiated blood cell lines. Finally, we compare the patterns we obtained with the characteristic emission patterns of ChromHMM, and show that ShallowChrome is able to coherently rank groups of chromatin states w.r.t. their transcriptional activity. CONCLUSIONS: In this work we demonstrate that it is possible to model HM-modulated gene expression regulation in a highly accurate, yet interpretable way. Our feature extraction algorithm leverages on data downstream the identification of enriched regions to retrieve gene-wise, statistically significant and dynamically located features for each HM. These features are highly predictive of gene transcriptional state, and allow for accurate modeling by computationally efficient logistic regression models. These models allow a direct inspection and a rigorous interpretation, helping to formulate quantifiable hypotheses.

BACKGROUND: Polyploidy and heterokaryosis are common and consequential genetic phenomena that increase the number of haplotypes in an organism and complicate whole-genome sequence analysis. Allele balance has been used to infer polyploidy and heterokaryosis in diverse organisms using read sets sequenced to greater than 50× whole-genome coverage. However, sequencing to adequate depth is costly if applied to multiple individuals or large genomes. RESULTS: We developed VCFvariance.pl to utilize the variance of allele balance to infer polyploidy and/or heterokaryosis at low sequence coverage. This analysis requires as little as 10× whole-genome coverage and reduces the allele balance profile down to a single value, which can be used to determine if an individual has two or more haplotypes. This approach was validated using simulated, synthetic, and authentic read sets from the oomycete species Bremia lactucae and Phytophthora infestans, the fungal species Saccharomyces cerevisiae, and the plant species Arabidopsis arenosa. This approach was deployed to determine that nine of 21 genotyped European race-type isolates of Bremia lactucae were inconsistent with diploidy and therefore likely heterokaryotic. CONCLUSIONS: Variance of allele balance is a reliable metric to detect departures from a diploid state, including polyploidy, heterokaryosis, a mixed sample, or chromosomal copy number variation. Deploying this strategy is computationally inexpensive, can reduce the cost of sequencing by up to 80%, and used to test any organism.

BACKGROUND: The widely spreading coronavirus disease (COVID-19) has three major spreading properties: pathogenic mutations, spatial, and temporal propagation patterns. We know the spread of the virus geographically and temporally in terms of statistics, i.e., the number of patients. However, we are yet to understand the spread at the level of individual patients. As of March 2021, COVID-19 is wide-spread all over the world with new genetic variants. One important question is to track the early spreading patterns of COVID-19 until the virus has got spread all over the world. RESULTS: In this work, we proposed AutoCoV, a deep learning method with multiple loss object, that can track the early spread of COVID-19 in terms of spatial and temporal patterns until the disease is fully spread over the world in July 2020. Performances in learning spatial or temporal patterns were measured with two clustering measures and one classification measure. For annotated SARS-CoV-2 sequences from the National Center for Biotechnology Information (NCBI), AutoCoV outperformed seven baseline methods in our experiments for learning either spatial or temporal patterns. For spatial patterns, AutoCoV had at least 1.7-fold higher clustering performances and an F1 score of 88.1%. For temporal patterns, AutoCoV had at least 1.6-fold higher clustering performances and an F1 score of 76.1%. Furthermore, AutoCoV demonstrated the robustness of the embedding space with an independent dataset, Global Initiative for Sharing All Influenza Data (GISAID). CONCLUSIONS: In summary, AutoCoV learns geographic and temporal spreading patterns successfully in experiments on NCBI and GISAID datasets and is the first of its kind that learns virus spreading patterns from the genome sequences, to the best of our knowledge. We expect that this type of embedding method will be helpful in characterizing fast-evolving pandemics.

BACKGROUND: SNARE proteins play an important role in different biological functions. This study aims to investigate the contribution of a new class of molecular descriptors (called SNARER) related to the chemical-physical properties of proteins in order to evaluate the performance of binary classifiers for SNARE proteins. RESULTS: We constructed a SNARE proteins balanced dataset, D128, and an unbalanced one, DUNI, on which we tested and compared the performance of the new descriptors presented here in combination with the feature sets (GAAC, CTDT, CKSAAP and 188D) already present in the literature. The machine learning algorithms used were Random Forest, k-Nearest Neighbors and AdaBoost and oversampling and subsampling techniques were applied to the unbalanced dataset. The addition of the SNARER descriptors increases the precision for all considered ML algorithms. In particular, on the unbalanced DUNI dataset the accuracy increases in parallel with the increase in sensitivity while on the balanced dataset D128 the accuracy increases compared to the counterpart without the addition of SNARER descriptors, with a strong improvement in specificity. Our best result is the combination of our descriptors SNARER with CKSAAP feature on the dataset D128 with 92.3% of accuracy, 90.1% for sensitivity and 95% for specificity with the RF algorithm. CONCLUSIONS: The performed analysis has shown how the introduction of molecular descriptors linked to the chemical-physical and structural characteristics of the proteins can improve the classification performance. Additionally, it was pointed out that performance can change based on using a balanced or unbalanced dataset. The balanced nature of training can significantly improve forecast accuracy.

BACKGROUND: High dimensional transcriptome profiling, whether through next generation sequencing techniques or high-throughput arrays, may result in scattered variables with missing data. Data imputation is a common strategy to maximize the inclusion of samples by using statistical techniques to fill in missing values. However, many data imputation methods are cumbersome and risk introduction of systematic bias. RESULTS: We present a new data imputation method using constrained least squares and algorithms from the inverse problems literature and present applications for this technique in miRNA expression analysis. The proposed technique is shown to offer an imputation orders of magnitude faster, with greater than or equal accuracy when compared to similar methods from the literature. CONCLUSIONS: This study offers a robust and efficient algorithm for data imputation, which can be used, e.g., to improve cancer prediction accuracy in the presence of missing data.

BACKGROUND: Over the past decade, experimental procedures such as metabolic labeling for determining RNA turnover rates at the transcriptome-wide scale have been widely adopted and are now turning to single cell measurements. Several computational methods to estimate RNA synthesis, processing and degradation rates from such experiments have been suggested, but they all require several RNA sequencing samples. Here we present a method that can estimate those three rates from a single sample. METHODS: Our method relies on the analytical solution to the Zeisel model of RNA dynamics. It was validated on metabolic labeling experiments performed on mouse embryonic stem cells. Resulting degradation rates were compared both to previously published rates on the same system and to a state-of-the-art method applied to the same data. RESULTS: Our method is computationally efficient and outputs rates that correlate well with previously published data sets. Using it on a single sample, we were able to reproduce the observation that dynamic biological processes tend to involve genes with higher metabolic rates, while stable processes involve genes with lower rates. This supports the hypothesis that cells control not only the mRNA steady-state abundance, but also its responsiveness, i.e., how fast steady state is reached. Moreover, degradation rates obtained with our method compare favourably with the other tested method. CONCLUSIONS: In addition to saving experimental work and computational time, estimating rates for a single sample has several advantages. It does not require an error-prone normalization across samples and enables the use of replicates to estimate uncertainty and assess sample quality. Finally the method and theoretical results described here are general enough to be useful in other contexts such as nucleotide conversion methods and single cell metabolic labeling experiments.

BACKGROUND: Autism spectrum disorder (ASD) is a group of complex neurodevelopment disorders with a strong genetic basis. Large scale sequencing studies have identified over one hundred ASD risk genes. Nevertheless, the vast majority of ASD risk genes remain to be discovered, as it is estimated that more than 1000 genes are likely to be involved in ASD risk. Prioritization of risk genes is an effective strategy to increase the power of identifying novel risk genes in genetics studies of ASD. As ASD risk genes are likely to exhibit distinct properties from multiple angles, we reason that integrating multiple levels of genomic data is a powerful approach to pinpoint genuine ASD risk genes. RESULTS: We present BNScore, a Bayesian model selection framework to probabilistically prioritize ASD risk genes through explicitly integrating evidence from sequencing-identified ASD genes, biological annotations, and gene functional network. We demonstrate the validity of our approach and its improved performance over existing methods by examining the resulting top candidate ASD risk genes against sets of high-confidence benchmark genes and large-scale ASD genome-wide association studies. We assess the tissue-, cell type- and development stage-specific expression properties of top prioritized genes, and find strong expression specificity in brain tissues, striatal medium spiny neurons, and fetal developmental stages. CONCLUSIONS: In summary, we show that by integrating sequencing findings, functional annotation profiles, and gene-gene functional network, our proposed BNScore provides competitive performance compared to current state-of-the-art methods in prioritizing ASD genes. Our method offers a general and flexible strategy to risk gene prioritization that can potentially be applied to other complex traits as well.

BACKGROUND: The abundance of biomedical text data coupled with advances in natural language processing (NLP) is resulting in novel biomedical NLP (BioNLP) applications. These NLP applications, or tasks, are reliant on the availability of domain-specific language models (LMs) that are trained on a massive amount of data. Most of the existing domain-specific LMs adopted bidirectional encoder representations from transformers (BERT) architecture which has limitations, and their generalizability is unproven as there is an absence of baseline results among common BioNLP tasks. RESULTS: We present 8 variants of BioALBERT, a domain-specific adaptation of a lite bidirectional encoder representations from transformers (ALBERT), trained on biomedical (PubMed and PubMed Central) and clinical (MIMIC-III) corpora and fine-tuned for 6 different tasks across 20 benchmark datasets. Experiments show that a large variant of BioALBERT trained on PubMed outperforms the state-of-the-art on named-entity recognition (+ 11.09% BLURB score improvement), relation extraction (+ 0.80% BLURB score), sentence similarity (+ 1.05% BLURB score), document classification (+ 0.62% F1-score), and question answering (+ 2.83% BLURB score). It represents a new state-of-the-art in 5 out of 6 benchmark BioNLP tasks. CONCLUSIONS: The large variant of BioALBERT trained on PubMed achieved a higher BLURB score than previous state-of-the-art models on 5 of the 6 benchmark BioNLP tasks. Depending on the task, 5 different variants of BioALBERT outperformed previous state-of-the-art models on 17 of the 20 benchmark datasets, showing that our model is robust and generalizable in the common BioNLP tasks. We have made BioALBERT freely available which will help the BioNLP community avoid computational cost of training and establish a new set of baselines for future efforts across a broad range of BioNLP tasks.

BACKGROUND: Despite recent progress in basecalling of Oxford nanopore DNA sequencing data, its wide adoption is still being hampered by its relatively low accuracy compared to short read technologies. Furthermore, very little of the recent research was focused on basecalling of RNA data, which has different characteristics than its DNA counterpart. RESULTS: We fill this gap by benchmarking a fully convolutional deep learning basecalling architecture with improved performance compared to Oxford nanopore's RNA basecallers. AVAILABILITY: The source code for our basecaller is available at: https://github.com/biodlab/RODAN .

De novo' drug discovery is costly, slow, and with high risk. Repurposing known drugs for treatment of other diseases offers a fast, low-cost/risk and highly-efficient method toward development of efficacious treatments. The emergence of large-scale heterogeneous biomolecular networks, molecular, chemical and bioactivity data, and genomic and phenotypic data of pharmacological compounds is enabling the development of new area of drug repurposing called 'in silico' drug repurposing, i.e., computational drug repurposing (CDR). The aim of CDR is to discover new indications for an existing drug (drug-centric) or to identify effective drugs for a disease (disease-centric). Both drug-centric and disease-centric approaches have the common challenge of either assessing the similarity or connections between drugs and diseases. However, traditional CDR is fraught with many challenges due to the underlying complex pharmacology and biology of diseases, genes, and drugs, as well as the complexity of their associations. As such, capturing highly non-linear associations among drugs, genes, diseases by most existing CDR methods has been challenging. We propose a network-based integration approach that can best capture knowledge (and complex relationships) contained within and between drugs, genes and disease data. A network-based machine learning approach is applied thereafter by using the extracted knowledge and relationships in order to identify single and pair of approved or experimental drugs with potential therapeutic effects on different breast cancer subtypes. Indeed, further clinical analysis is needed to confirm the therapeutic effects of identified drugs on each breast cancer subtype.

BACKGROUND: Epistasis is the interaction between different genes when expressing a certain phenotype. If epistasis involves more than two loci it is called high-order epistasis. High-order epistasis is an area under active research because it could be the cause of many complex traits. The most common way to specify an epistasis interaction is through a penetrance table. RESULTS: This paper presents PyToxo, a Python tool for generating penetrance tables from any-order epistasis models. Unlike other tools available in the bibliography, PyToxo is able to work with high-order models and realistic penetrance and heritability values, achieving high-precision results in a short time. In addition, PyToxo is distributed as open-source software and includes several interfaces to ease its use. CONCLUSIONS: PyToxo provides the scientific community with a useful tool to evaluate algorithms and methods that can detect high-order epistasis to continue advancing in the discovery of the causes behind complex diseases.

MOTIVATION: Deep learning has become a prevalent method in identifying genomic regulatory sequences such as promoters. In a number of recent papers, the performance of deep learning models has continually been reported as an improvement over alternatives for sequence-based promoter recognition. However, the performance improvements in these models do not account for the different datasets that models are evaluated on. The lack of a consensus dataset and procedure for benchmarking purposes has made the comparison of each model's true performance difficult to assess. RESULTS: We present a framework called Supervised Promoter Recognition Framework ('SUPR REF') capable of streamlining the complete process of training, validating, testing, and comparing promoter recognition models in a systematic manner. SUPR REF includes the creation of biologically relevant benchmark datasets to be used in the evaluation process of deep learning promoter recognition models. We showcase this framework by comparing the models' performances on alternative datasets, and properly evaluate previously published models on new benchmark datasets. Our results show that the reliability of deep learning ab initio promoter recognition models on eukaryotic genomic sequences is still not at a sufficient level, as overall performance is still low. These results originate from a subset of promoters, the well-known RNA Polymerase II core promoters. Furthermore, given the observational nature of these data, cross-validation results from small promoter datasets need to be interpreted with caution.

BACKGROUND: Understanding the role of various factors in 3D genome organization is essential to determine their impact on shaping large-scale chromatin units such as euchromatin (A) and heterochromatin (B) compartments. At this level, chromatin compaction is extensively modulated when transcription and epigenetic profiles change upon cell differentiation and response to various external impacts. However, detailed analysis of chromatin contact patterns within and between compartments is complicated because of a lack of suitable computational methods. RESULTS: We developed a tool, Pentad, to perform calculation, visualisation and quantitative analysis of the average chromatin compartment from the Hi-C matrices in cis, trans, and specified genomic distances. As we demonstrated by applying Pentad to publicly available Hi-C datasets, it helps to reliably detect redistribution of contact frequency in the chromatin compartments and assess alterations in the compartment strength. CONCLUSIONS: Pentad is a simple tool for the analysis of changes in chromatin compartmentalization in various biological conditions. Pentad is freely available at https://github.com/magnitov/pentad .

BACKGROUND: Estimation of the accuracy (quality) of protein structural models is important for both prediction and use of protein structural models. Deep learning methods have been used to integrate protein structure features to predict the quality of protein models. Inter-residue distances are key information for predicting protein's tertiary structures and therefore have good potentials to predict the quality of protein structural models. However, few methods have been developed to fully take advantage of predicted inter-residue distance maps to estimate the accuracy of a single protein structural model. RESULT: We developed an attentive 2D convolutional neural network (CNN) with channel-wise attention to take only a raw difference map between the inter-residue distance map calculated from a single protein model and the distance map predicted from the protein sequence as input to predict the quality of the model. The network comprises multiple convolutional layers, batch normalization layers, dense layers, and Squeeze-and-Excitation blocks with attention to automatically extract features relevant to protein model quality from the raw input without using any expert-curated features. We evaluated DISTEMA's capability of selecting the best models for CASP13 targets in terms of ranking loss of GDT-TS score. The ranking loss of DISTEMA is 0.079, lower than several state-of-the-art single-model quality assessment methods. CONCLUSION: This work demonstrates that using raw inter-residue distance information with deep learning can predict the quality of protein structural models reasonably well. DISTEMA is freely at https://github.com/jianlin-cheng/DISTEMA.

BACKGROUND: The gene signatures have been considered as a promising early diagnosis and prognostic analysis to identify disease subtypes and to determine subsequent treatments. Tissue-specific gene signatures of a specific disease are an emergency requirement for precision medicine to improve the accuracy and reduce the side effects. Currently, many approaches have been proposed for identifying gene signatures for diagnosis and prognostic. However, they often lack of tissue-specific gene signatures. RESULTS: Here, we propose a new method, consensus mutual information (CoMI) for analyzing omics data and discovering gene signatures. CoMI can identify differentially expressed genes in multiple cancer omics data for reflecting both cancer-related and tissue-specific signatures, such as Cell growth and death in multiple cancers, Xenobiotics biodegradation and metabolism in LIHC, and Nervous system in GBM. Our method identified 50-gene signatures effectively distinguishing the GBM patients into high- and low-risk groups (log-rank p = 0.006) for diagnosis and prognosis. CONCLUSIONS: Our results demonstrate that CoMI can identify significant and consistent gene signatures with tissue-specific properties and can predict clinical outcomes for interested diseases. We believe that CoMI is useful for analyzing omics data and discovering gene signatures of diseases.

BACKGROUND: Chronic cough affects approximately 10% of adults. The lack of ICD codes for chronic cough makes it challenging to apply supervised learning methods to predict the characteristics of chronic cough patients, thereby requiring the identification of chronic cough patients by other mechanisms. We developed a deep clustering algorithm with auto-encoder embedding (DCAE) to identify clusters of chronic cough patients based on data from a large cohort of 264,146 patients from the Electronic Medical Records (EMR) system. We constructed features using the diagnosis within the EMR, then built a clustering-oriented loss function directly on embedded features of the deep autoencoder to jointly perform feature refinement and cluster assignment. Lastly, we performed statistical analysis on the identified clusters to characterize the chronic cough patients compared to the non-chronic cough patients. RESULTS: The experimental results show that the DCAE model generated three chronic cough clusters and one non-chronic cough patient cluster. We found various diagnoses, medications, and lab tests highly associated with chronic cough patients by comparing the chronic cough cluster with the non-chronic cough cluster. Comparison of chronic cough clusters demonstrated that certain combinations of medications and diagnoses characterize some chronic cough clusters. CONCLUSIONS: To the best of our knowledge, this study is the first to test the potential of unsupervised deep learning methods for chronic cough investigation, which also shows a great advantage over existing algorithms for patient data clustering.

BACKGROUND: With a growing amount of (multi-)omics data being available, the extraction of knowledge from these datasets is still a difficult problem. Classical enrichment-style analyses require predefined pathways or gene sets that are tested for significant deregulation to assess whether the pathway is functionally involved in the biological process under study. De novo identification of these pathways can reduce the bias inherent in predefined pathways or gene sets. At the same time, the definition and efficient identification of these pathways de novo from large biological networks is a challenging problem. RESULTS: We present a novel algorithm, DeRegNet, for the identification of maximally deregulated subnetworks on directed graphs based on deregulation scores derived from (multi-)omics data. DeRegNet can be interpreted as maximum likelihood estimation given a certain probabilistic model for de-novo subgraph identification. We use fractional integer programming to solve the resulting combinatorial optimization problem. We can show that the approach outperforms related algorithms on simulated data with known ground truths. On a publicly available liver cancer dataset we can show that DeRegNet can identify biologically meaningful subgraphs suitable for patient stratification. DeRegNet can also be used to find explicitly multi-omics subgraphs which we demonstrate by presenting subgraphs with consistent methylation-transcription patterns. DeRegNet is freely available as open-source software. CONCLUSION: The proposed algorithmic framework and its available implementation can serve as a valuable heuristic hypothesis generation tool contextualizing omics data within biomolecular networks.

BACKGROUND: Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths worldwide. Recent studies have observed causative mutations in susceptible genes related to colorectal cancer in 10 to 15% of the patients. This highlights the importance of identifying mutations for early detection of this cancer for more effective treatments among high risk individuals. Mutation is considered as the key point in cancer research. Many studies have performed cancer subtyping based on the type of frequently mutated genes, or the proportion of mutational processes. However, to the best of our knowledge, combination of these features has never been used together for this task. This highlights the potential to introduce better and more inclusive subtype classification approaches using wider range of related features to enable biomarker discovery and thus inform drug development for CRC. RESULTS: In this study, we develop a new pipeline based on a novel concept called 'gene-motif', which merges mutated gene information with tri-nucleotide motif of mutated sites, for colorectal cancer subtype identification. We apply our pipeline to the International Cancer Genome Consortium (ICGC) CRC samples and identify, for the first time, 3131 gene-motif combinations that are significantly mutated in 536 ICGC colorectal cancer samples. Using these features, we identify seven CRC subtypes with distinguishable phenotypes and biomarkers, including unique cancer related signaling pathways, in which for most of them targeted treatment options are currently available. Interestingly, we also identify several genes that are mutated in multiple subtypes but with unique sequence contexts. CONCLUSION: Our results highlight the importance of considering both the mutation type and mutated genes in identification of cancer subtypes and cancer biomarkers. The new CRC subtypes presented in this study demonstrates distinguished phenotypic properties which can be effectively used to develop new treatments. By knowing the genes and phenotypes associated with the subtypes, a personalized treatment plan can be developed that considers the specific phenotypes associated with their genomic lesion.

BACKGROUND: Being able to efficiently call variants from the increasing amount of sequencing data daily produced from multiple viral strains is of the utmost importance, as demonstrated during the COVID-19 pandemic, in order to track the spread of the viral strains across the globe. RESULTS: We present MALVIRUS, an easy-to-install and easy-to-use application that assists users in multiple tasks required for the analysis of a viral population, such as the SARS-CoV-2. MALVIRUS allows to: (1) construct a variant catalog consisting in a set of variations (SNPs/indels) from the population sequences, (2) efficiently genotype and annotate variants of the catalog supported by a read sample, and (3) when the considered viral species is the SARS-CoV-2, assign the input sample to the most likely Pango lineages using the genotyped variations. CONCLUSIONS: Tests on Illumina and Nanopore samples proved the efficiency and the effectiveness of MALVIRUS in analyzing SARS-CoV-2 strain samples with respect to publicly available data provided by NCBI and the more complete dataset provided by GISAID. A comparison with state-of-the-art tools showed that MALVIRUS is always more precise and often have a better recall.

BACKGROUND: SARS-CoV-2 virus sequencing has been applied to track the COVID-19 pandemic spread and assist the development of PCR-based diagnostics, serological assays, and vaccines. With sequencing becoming routine globally, bioinformatic tools are needed to assist in the robust processing of resulting genomic data. RESULTS: We developed a web-based bioinformatic pipeline ("COVID-Profiler") that inputs raw or assembled sequencing data, displays raw alignments for quality control, annotates mutations found and performs phylogenetic analysis. The pipeline software can be applied to other (re-) emerging pathogens. CONCLUSIONS: The webserver is available at http://genomics.lshtm.ac.uk/ . The source code is available at https://github.com/jodyphelan/covid-profiler .

BACKGROUND: Mass spectrometry imaging (MSI) data often consist of tens of thousands of mass spectra collected from a sample surface. During the time necessary to perform a single acquisition, it is likely that uncontrollable factors alter the validity of the initial mass calibration of the instrument, resulting in mass errors of magnitude significantly larger than their theoretical values. This phenomenon has a two-fold detrimental effect: (a) it reduces the ability to interpret the results based on the observed signals, (b) it can affect the quality of the observed signal spatial distributions. RESULTS: We present a post-acquisition computational method capable of reducing the observed mass drift by up to 60 ppm in biological samples, exploiting the presence of typical molecules with a known mass-to-charge ratio. The procedure, tested on time-of-flight and Orbitrap mass spectrometry analyzers interfaced to a desorption electrospray ionization (DESI) source, improves the molecular annotation quality and the spatial distributions of the detected ions. CONCLUSION: The presented method represents a robust and accurate tool for performing post-acquisition mass recalibration of DESI-MSI datasets and can help to increase the reliability of the molecular assignment and the data quality.

BACKGROUND: Drug resistance is a critical obstacle in cancer therapy. Discovering cancer drug response is important to improve anti-cancer drug treatment and guide anti-cancer drug design. Abundant genomic and drug response resources of cancer cell lines provide unprecedented opportunities for such study. However, cancer cell lines cannot fully reflect heterogeneous tumor microenvironments. Transferring knowledge studied from in vitro cell lines to single-cell and clinical data will be a promising direction to better understand drug resistance. Most current studies include single nucleotide variants (SNV) as features and focus on improving predictive ability of cancer drug response on cell lines. However, obtaining accurate SNVs from clinical tumor samples and single-cell data is not reliable. This makes it difficult to generalize such SNV-based models to clinical tumor data or single-cell level studies in the future. RESULTS: We present a new method, DualGCN, a unified Dual Graph Convolutional Network model to predict cancer drug response. DualGCN encodes both chemical structures of drugs and omics data of biological samples using graph convolutional networks. Then the two embeddings are fed into a multilayer perceptron to predict drug response. DualGCN incorporates prior knowledge on cancer-related genes and protein-protein interactions, and outperforms most state-of-the-art methods while avoiding using large-scale SNV data. CONCLUSIONS: The proposed method outperforms most state-of-the-art methods in predicting cancer drug response without the use of large-scale SNV data. These favorable results indicate its potential to be extended to clinical and single-cell tumor samples and advancements in precision medicine.

BACKGROUND: Predicting which pathogens might exhibit antimicrobial resistance (AMR) based on genomics data is one of the promising ways to swiftly and precisely identify AMR pathogens. Currently, the most widely used genomics approach is through identifying known AMR genes from genomic information in order to predict whether a pathogen might be resistant to certain antibiotic drugs. The list of known AMR genes, however, is still far from comprehensive and may result in inaccurate AMR pathogen predictions. We thus felt the need to expand the AMR gene set and proposed a pan-genome-based feature selection method to identify potential gene sets for AMR prediction purposes. RESULTS: By building pan-genome datasets and extracting gene presence/absence patterns from four bacterial species, each with more than 2000 strains, we showed that machine learning models built from pan-genome data can be very promising for predicting AMR pathogens. The gene set selected by the eXtreme Gradient Boosting (XGBoost) feature selection approach further improved prediction outcomes, and an incremental approach selecting subsets of XGBoost-selected features brought the machine learning model performance to the next level. Investigating selected gene sets revealed that on average about 50% of genes had no known function and very few of them were known AMR genes, indicating the potential of the selected gene sets to expand resistance gene repertoires. CONCLUSIONS: We demonstrated that a pan-genome-based feature selection approach is suitable for building machine learning models for predicting AMR pathogens. The extracted gene sets may provide future clues to expand our knowledge of known AMR genes and provide novel hypotheses for inferring bacterial AMR mechanisms.

BACKGROUND: Medical information has rapidly increased on the internet and has become one of the main targets of search engine use. However, medical information on the internet is subject to the problems of quality and accessibility, so ordinary users are unable to obtain answers to their medical questions conveniently. As a solution, researchers build medical question answering (QA) systems. However, research on medical QA in the Chinese language lags behind work on English-based systems. This lag is mainly due to the difficulty of constructing a high-quality knowledge base and the underutilization of medical corpora in the Chinese language. RESULTS: This study developed an end-to-end solution to implement a medical QA system for the Chinese language with low cost and time. First, we created a high-quality medical knowledge graph from hospital data (electronic health/medical records) in a nearly automatic manner that trained a supervised model based on data labeled using bootstrapping techniques. Then, we designed a QA system based on a memory-based neural network and attention mechanism. Finally, we trained the system to generate answers from the knowledge base and a QA corpus on the internet. CONCLUSIONS: Bootstrapping and deep neural network techniques can construct a knowledge graph from electronic health/medical records with satisfactory precision and coverage. Our proposed context bridge mechanisms perform training with a variety of language features. Our QA system can achieve state-of-the-art quality in answering medical questions with constrained topics. As we evaluated, complex Chinese language processing techniques, such as segmentation and parsing, were not necessary for practice and complex architectures were not necessary to build the QA system. Lastly, we created an application using our method for internet QA usage.

BACKGROUND: Retroviruses replicate by integrating a DNA copy into a host chromosome. Detecting novel retroviral integrations (ones not in the reference genome sequence of the host) from genomic NGS data is bioinformatically challenging and frequently produces many false positives. One common method of confirmation is visual inspection of an alignment of the chimaeric (split) reads that span a putative novel retroviral integration site. We perceived the need for a program that would facilitate this by producing a multiple alignment containing both the viral and host regions that flank an integration. RESULTS: BreakAlign is a Perl program that uses blastn to produce such a multiple alignment. In addition to the NGS dataset and a reference viral sequence, the program requires either (a) the ~ 500nt host genome sequence that spans the putative integration or (b) coordinates of this putative integration in an installed copy of the reference human genome (multiple integrations can be processed automatically). BreakAlign is freely available from https://github.com/marchiem/breakalign and is accompanied by example files allowing a test run. CONCLUSION: BreakAlign will confirm and facilitate characterisation of both (a) germline integrations of endogenous retroviruses and (b) somatic integrations of exogenous retroviruses such as HIV and HTLV. Although developed for use with genomic short-read NGS (second generation) data and retroviruses, it should also be useful for long-read (third generation) data and any mobile element with at least one conserved flanking region.

BACKGROUND: Human protein kinases play important roles in cancers, are highly co-regulated by kinase families rather than a single kinase, and complementarily regulate signaling pathways. Even though there are > 100,000 protein kinase inhibitors, only 67 kinase drugs are currently approved by the Food and Drug Administration (FDA). RESULTS: In this study, we used "merged moiety-based interpretable features (MMIFs)," which merged four moiety-based compound features, including Checkmol fingerprint, PubChem fingerprint, rings in drugs, and in-house moieties as the input features for building random forest (RF) models. By using > 200,000 bioactivity test data, we classified inhibitors as kinase family inhibitors or non-inhibitors in the machine learning. The results showed that our RF models achieved good accuracy (> 0.8) for the 10 kinase families. In addition, we found kinase common and specific moieties across families using the Shapley Additive exPlanations (SHAP) approach. We also verified our results using protein kinase complex structures containing important interactions of the hinges, DFGs, or P-loops in the ATP pocket of active sites. CONCLUSIONS: In summary, we not only constructed highly accurate prediction models for predicting inhibitors of kinase families but also discovered common and specific inhibitor moieties between different kinase families, providing new opportunities for designing protein kinase inhibitors.

BACKGROUND: Converting molecules into computer-interpretable features with rich molecular information is a core problem of data-driven machine learning applications in chemical and drug-related tasks. Generally speaking, there are global and local features to represent a given molecule. As most algorithms have been developed based on one type of feature, a remaining bottleneck is to combine both feature sets for advanced molecule-based machine learning analysis. Here, we explored a novel analytical framework to make embeddings of the molecular features and apply them in the clustering of a large number of small molecules. RESULTS: In this novel framework, we first introduced a principal component analysis method encoding the molecule-specific atom and bond information. We then used a variational autoencoder (AE)-based method to make embeddings of the global chemical properties and the local atom and bond features. Next, using the embeddings from the encoded local and global features, we implemented and compared several unsupervised clustering algorithms to group the molecule-specific embeddings. The number of clusters was treated as a hyper-parameter and determined by the Silhouette method. Finally, we evaluated the corresponding results using three internal indices. Applying the analysis framework to a large chemical library of more than 47,000 molecules, we successfully identified 50 molecular clusters using the K-means method with 32 embeddings based on the AE method. We visualized the clustering result via t-SNE for the overall distribution of molecules and the similarity maps for the structural analysis of randomly selected cluster-specific molecules. CONCLUSIONS: This study developed a novel analytical framework that comprises a feature engineering scheme for molecule-specific atomic and bonding features and a deep learning-based embedding strategy for different molecular features. By applying the identified embeddings, we show their usefulness for clustering a large molecule dataset. Our novel analytic algorithms can be applied to any virtual library of chemical compounds with diverse molecular structures. Hence, these tools have the potential of optimizing drug discovery, as they can decrease the number of compounds to be screened in any drug screening campaign.

BACKGROUND: Long non-coding RNA (LncRNA) plays important roles in physiological and pathological processes. Identifying LncRNA-protein interactions (LPIs) is essential to understand the molecular mechanism and infer the functions of lncRNAs. With the overwhelming size of the biomedical literature, extracting LPIs directly from the biomedical literature is essential, promising and challenging. However, there is no webserver of LPIs relationship extraction from literature. RESULTS: LPInsider is developed as the first webserver for extracting LPIs from biomedical literature texts based on multiple text features (semantic word vectors, syntactic structure vectors, distance vectors, and part of speech vectors) and logistic regression. LPInsider allows researchers to extract LPIs by uploading PMID, PMCID, PMID List, or biomedical text. A manually filtered and highly reliable LPI corpus is integrated in LPInsider. The performance of LPInsider is optimal by comprehensive experiment on different combinations of different feature and machine learning models. CONCLUSIONS: LPInsider is an efficient analytical tool for LPIs that helps researchers to enhance their comprehension of lncRNAs from text mining, and also saving their time. In addition, LPInsider is freely accessible from http://www.csbg-jlu.info/LPInsider/ with no login requirement. The source code and LPIs corpus can be downloaded from https://github.com/qiufengdiewu/LPInsider .

BACKGROUND: Topologically associating domains (TADs) are locally highly-interacting genome regions, which also play a critical role in regulating gene expression in the cell. TADs have been first identified while investigating the 3D genome structure over High-throughput Chromosome Conformation Capture (Hi-C) interaction dataset. Substantial degree of efforts have been devoted to develop techniques for inferring TADs from Hi-C interaction dataset. Many TAD-calling methods have been developed which differ in their criteria and assumptions in TAD inference. Correspondingly, TADs inferred via these callers vary in terms of both similarities and biological features they are enriched in. RESULT: We have carried out a systematic comparison of 27 TAD-calling methods over mammals. We use Micro-C, a recent high-resolution variant of Hi-C, to compare TADs at a very high resolution, and classify the methods into 3 categories: feature-based methods, Clustering methods, Graph-partitioning methods. We have evaluated TAD boundaries, gaps between adjacent TADs, and quality of TADs across various criteria. We also found particularly CTCF and Cohesin proteins to be effective in formation of TADs with corner dots. We have also assessed the callers performance on simulated datasets since a gold standard for TADs is missing. TAD sizes and numbers change remarkably between TAD callers and dataset resolutions, indicating that TADs are hierarchically-organized domains, instead of disjoint regions. A core subset of feature-based TAD callers regularly perform the best while inferring reproducible domains, which are also enriched for TAD related biological properties. CONCLUSION: We have analyzed the fundamental principles of TAD-calling methods, and identified the existing situation in TAD inference across high resolution Micro-C interaction datasets over mammals. We come up with a systematic, comprehensive, and concise framework to evaluate the TAD-calling methods performance across Micro-C datasets. Our research will be useful in selecting appropriate methods for TAD inference and evaluation based on available data, experimental design, and biological question of interest. We also introduce our analysis as a benchmarking tool with publicly available source code.

BACKGROUND: With the development of noninvasive imaging technology, collecting different imaging measurements of the same brain has become more and more easy. These multimodal imaging data carry complementary information of the same brain, with both specific and shared information being intertwined. Within these multimodal data, it is essential to discriminate the specific information from the shared information since it is of benefit to comprehensively characterize brain diseases. While most existing methods are unqualified, in this paper, we propose a parameter decomposition based sparse multi-view canonical correlation analysis (PDSMCCA) method. PDSMCCA could identify both modality-shared and -specific information of multimodal data, leading to an in-depth understanding of complex pathology of brain disease. RESULTS: Compared with the SMCCA method, our method obtains higher correlation coefficients and better canonical weights on both synthetic data and real neuroimaging data. This indicates that, coupled with modality-shared and -specific feature selection, PDSMCCA improves the multi-view association identification and shows meaningful feature selection capability with desirable interpretation. CONCLUSIONS: The novel PDSMCCA confirms that the parameter decomposition is a suitable strategy to identify both modality-shared and -specific imaging features. The multimodal association and the diverse information of multimodal imaging data enable us to better understand the brain disease such as Alzheimer's disease.

BACKGROUND: In research on new drug discovery, the traditional wet experiment has a long period. Predicting drug-target interaction (DTI) in silico can greatly narrow the scope of search of candidate medications. Excellent algorithm model may be more effective in revealing the potential connection between drug and target in the bioinformatics network composed of drugs, proteins and other related data. RESULTS: In this work, we have developed a heterogeneous graph neural network model, named as HGDTI, which includes a learning phase of network node embedding and a training phase of DTI classification. This method first obtains the molecular fingerprint information of drugs and the pseudo amino acid composition information of proteins, then extracts the initial features of nodes through Bi-LSTM, and uses the attention mechanism to aggregate heterogeneous neighbors. In several comparative experiments, the overall performance of HGDTI significantly outperforms other state-of-the-art DTI prediction models, and the negative sampling technology is employed to further optimize the prediction power of model. In addition, we have proved the robustness of HGDTI through heterogeneous network content reduction tests, and proved the rationality of HGDTI through other comparative experiments. These results indicate that HGDTI can utilize heterogeneous information to capture the embedding of drugs and targets, and provide assistance for drug development. CONCLUSIONS: The HGDTI based on heterogeneous graph neural network model, can utilize heterogeneous information to capture the embedding of drugs and targets, and provide assistance for drug development. For the convenience of related researchers, a user-friendly web-server has been established at http://bioinfo.jcu.edu.cn/hgdti .

BACKGROUND: Electronic medical records (EMR) contain detailed information about patient health. Developing an effective representation model is of great significance for the downstream applications of EMR. However, processing data directly is difficult because EMR data has such characteristics as incompleteness, unstructure and redundancy. Therefore, preprocess of the original data is the key step of EMR data mining. The classic distributed word representations ignore the geometric feature of the word vectors for the representation of EMR data, which often underestimate the similarities between similar words and overestimate the similarities between distant words. This results in word similarity obtained from embedding models being inconsistent with human judgment and much valuable medical information being lost. RESULTS: In this study, we propose a biomedical word embedding framework based on manifold subspace. Our proposed model first obtains the word vector representations of the EMR data, and then re-embeds the word vector in the manifold subspace. We develop an efficient optimization algorithm with neighborhood preserving embedding based on manifold optimization. To verify the algorithm presented in this study, we perform experiments on intrinsic evaluation and external classification tasks, and the experimental results demonstrate its advantages over other baseline methods. CONCLUSIONS: Manifold learning subspace embedding can enhance the representation of distributed word representations in electronic medical record texts. Reduce the difficulty for researchers to process unstructured electronic medical record text data, which has certain biomedical research value.