BACKGROUND: Drug-drug interactions (DDIs) are the reactions between drugs. They are compartmentalized into three types: synergistic, antagonistic and no reaction. As a rapidly developing technology, predicting DDIs-associated events is getting more and more attention and application in drug development and disease diagnosis fields. In this work, we study not only whether the two drugs interact, but also specific interaction types. And we propose a learning-based method using convolution neural networks to learn feature representations and predict DDIs. RESULTS: In this paper, we proposed a novel algorithm using a CNN architecture, named CNN-DDI, to predict drug-drug interactions. First, we extract feature interactions from drug categories, targets, pathways and enzymes as feature vectors and employ the Jaccard similarity as the measurement of drugs similarity. Then, based on the representation of features, we build a new convolution neural network as the DDIs' predictor. CONCLUSION: The experimental results indicate that drug categories is effective as a new feature type applied to CNN-DDI method. And using multiple features is more informative and more effective than single feature. It can be concluded that CNN-DDI has more superiority than other existing algorithms on task of predicting DDIs.

BACKGROUND: Measuring similarity between complex diseases has significant implications for revealing the pathogenesis of diseases and development in the domain of biomedicine. It has been consentaneous that functional associations between disease-related genes and semantic associations can be applied to calculate disease similarity. Currently, more and more studies have demonstrated the profound involvement of non-coding RNA in the regulation of genome organization and gene expression. Thus, taking ncRNA into account can be useful in measuring disease similarities. However, existing methods ignore the regulation functions of ncRNA in biological process. In this study, we proposed a novel deep-learning method to deduce disease similarity. RESULTS: In this article, we proposed a novel method, ImpAESim, a framework integrating multiple networks embedding to learn compact feature representations and disease similarity calculation. We first utilize three different disease-related information networks to build up a heterogeneous network, after a network diffusion process, RWR, a compact feature learning model composed of classic Auto Encoder (AE) and improved AE model is proposed to extract constraints and low-dimensional feature representations. We finally obtain an accurate and low-dimensional feature representation of diseases, then we employed the cosine distance as the measurement of disease similarity. CONCLUSION: ImpAESim focuses on extracting a low-dimensional vector representation of features based on ncRNA regulation, and gene-gene interaction network. Our method can significantly reduce the calculation bias resulted from the sparse disease associations which are derived from semantic associations.

BACKGROUND: Locomotive behaviors are a rapid evaluation indicator reflecting whether the nervous system of worms is damaged, and has been proved to be sensitive to chemical toxicity. In many toxicological studies, C. elegans head thrashes is a key indicator of locomotive behaviors to measure the vitality of worms. In previous studies, the number of head thrashes was manually counted, which is time-consuming and labor-intensive. RESULTS: This paper presents an automatic recognition and counting method for head thrashes behavior of worms from experimental videos. First, the image processing algorithm is designed for worm morphology features calculation, mean gray values of head and tail are used to locate the head of worm accurately. Next, the worm skeleton is extracted and divided into equal parts. The angle formulas are used to calculate the bending angle of the head of worm. Finally, the number of head thrashes is counted according to the bending angle of the head in each frame. The robustness of the proposed algorithm is evaluated by comparing the counting results of the manual counting. It is proved that the proposed algorithm can recognize the occurrence of head thrashes of C. elegans of different strains. In addition, the difference of the head thrashes behavior of different worm strains is analyzed, it is proved that the relationship between worm head thrashes behavior and lifespan. CONCLUSIONS: A new method is proposed to automatically count the number of head thrashes of worms. This algorithm makes it possible to count the number of head thrashes from the worm videos collected by the automatic tracking system. The proposed algorithm will play an important role in toxicological research and worm vitality research. The code is freely available at https://github.com/hthana/HTC .

BACKGROUND: To date, cancer still is one of the leading causes of death worldwide, in which the cumulative of genes carrying mutations was said to be held accountable for the establishment and development of this disease mainly. From that, identification and analysis of driver genes were vital. Our previous study indicated disagreement on a unifying pipeline for these tasks and then introduced a complete one. However, this pipeline gradually manifested its weaknesses as being unfamiliar to non-technical users, time-consuming, and inconvenient. RESULTS: This study presented an R package named DrGA, developed based on our previous pipeline, to tackle the mentioned problems above. It wholly automated four widely used downstream analyses for predicted driver genes and offered additional improvements. We described the usage of the DrGA on driver genes of human breast cancer. Besides, we also gave the users another potential application of DrGA in analyzing genomic biomarkers of a complex disease in another organism. CONCLUSIONS: DrGA facilitated the users with limited IT backgrounds and rapidly created consistent and reproducible results. DrGA and its applications, along with example data, were freely provided at https://github.com/huynguyen250896/DrGA .

BACKGROUND: A typical Copy Number Variations (CNVs) detection process based on the depth of coverage in the Whole Exome Sequencing (WES) data consists of several steps: (I) calculating the depth of coverage in sequencing regions, (II) quality control, (III) normalizing the depth of coverage, (IV) calling CNVs. Previous tools performed one normalization process for each chromosome-all the coverage depths in the sequencing regions from a given chromosome were normalized in a single run. METHODS: Herein, we present the new CNVind tool for calling CNVs, where the normalization process is conducted separately for each of the sequencing regions. The total number of normalizations is equal to the number of sequencing regions in the investigated dataset. For example, when analyzing a dataset composed of n sequencing regions, CNVind performs n independent depth of coverage normalizations. Before each normalization, the application selects the k most correlated sequencing regions with the depth of coverage Pearson's Correlation as distance metric. Then, the resulting subgroup of [Formula: see text] sequencing regions is normalized, the results of all n independent normalizations are combined; finally, the segmentation and CNV calling process is performed on the resultant dataset. RESULTS AND CONCLUSIONS: We used WES data from the 1000 Genomes project to evaluate the impact of independent normalization on CNV calling performance and compared the results with state-of-the-art tools: CODEX and exomeCopy. The results proved that independent normalization allows to improve the rare CNVs detection specificity significantly. For example, for the investigated dataset, we reduced the number of FP calls from over 15,000 to around 5000 while maintaining a constant number of TP calls equal to about 150 CNVs. However, independent normalization of each sequencing region is a computationally expensive process, therefore our pipeline is customized and can be easily run in the cloud computing environment, on the computer cluster, or the single CPU server. To our knowledge, the presented application is the first attempt to implement an innovative approach to independent normalization of the depth of WES data coverage.

BACKGROUND: Drug repurposing aims at finding new targets for already developed drugs. It becomes more relevant as the cost of discovering new drugs steadily increases. To find new potential targets for a drug, an abundance of methods and existing biomedical knowledge from different domains can be leveraged. Recently, knowledge graphs have emerged in the biomedical domain that integrate information about genes, drugs, diseases and other biological domains. Knowledge graphs can be used to predict new connections between compounds and diseases, leveraging the interconnected biomedical data around them. While real world use cases such as drug repurposing are only interested in one specific relation type, widely used knowledge graph embedding models simultaneously optimize over all relation types in the graph. This can lead the models to underfit the data that is most relevant for the desired relation type. For example, if we want to learn embeddings to predict links between compounds and diseases but almost the entirety of relations in the graph is incident to other pairs of entity types, then the resulting embeddings are likely not optimised to predict links between compounds and diseases. We propose a method that leverages domain knowledge in the form of metapaths and use them to filter two biomedical knowledge graphs (Hetionet and DRKG) for the purpose of improving performance on the prediction task of drug repurposing while simultaneously increasing computational efficiency. RESULTS: We find that our method reduces the number of entities by 60% on Hetionet and 26% on DRKG, while leading to an improvement in prediction performance of up to 40.8% on Hetionet and 14.2% on DRKG, with an average improvement of 20.6% on Hetionet and 8.9% on DRKG. Additionally, prioritization of antiviral compounds for SARS CoV-2 improves after task-driven filtering is applied. CONCLUSION: Knowledge graphs contain facts that are counter productive for specific tasks, in our case drug repurposing. We also demonstrate that these facts can be removed, resulting in an improved performance in that task and a more efficient learning process.

BACKGROUND: Databases are fundamental to advance biomedical science. However, most of them are populated and updated with a great deal of human effort. Biomedical Relation Extraction (BioRE) aims to shift this burden to machines. Among its different applications, the discovery of Gene-Disease Associations (GDAs) is one of BioRE most relevant tasks. Nevertheless, few resources have been developed to train models for GDA extraction. Besides, these resources are all limited in size-preventing models from scaling effectively to large amounts of data. RESULTS: To overcome this limitation, we have exploited the DisGeNET database to build a large-scale, semi-automatically annotated dataset for GDA extraction. DisGeNET stores one of the largest available collections of genes and variants involved in human diseases. Relying on DisGeNET, we developed TBGA: a GDA extraction dataset generated from more than 700K publications that consists of over 200K instances and 100K gene-disease pairs. Each instance consists of the sentence from which the GDA was extracted, the corresponding GDA, and the information about the gene-disease pair. CONCLUSIONS: TBGA is amongst the largest datasets for GDA extraction. We have evaluated state-of-the-art models for GDA extraction on TBGA, showing that it is a challenging and well-suited dataset for the task. We made the dataset publicly available to foster the development of state-of-the-art BioRE models for GDA extraction.

BACKGROUND: Microalgae are emerging as promising sustainable sources for biofuels, biostimulants in agriculture, soil bioremediation, feed and human nutrients. Nonetheless, the molecular mechanisms underpinning microalgae physiology and the biosynthesis of compounds of biotechnological interest are largely uncharacterized. This hinders the development of microalgae full potential as cell-factories. The recent application of omics technologies into microalgae research aims at unraveling these systems. Nevertheless, the lack of specific tools for analysing omics raw data generated from microalgae to provide biological meaningful information are hampering the impact of these technologies. The purpose of ALGAEFUN with MARACAS consists in providing researchers in microalgae with an enabling tool that will allow them to exploit transcriptomic and cistromic high-throughput sequencing data. RESULTS: ALGAEFUN with MARACAS consists of two different tools. First, MARACAS (MicroAlgae RnA-seq and Chip-seq AnalysiS) implements a fully automatic computational pipeline receiving as input RNA-seq (RNA sequencing) or ChIP-seq (chromatin immunoprecipitation sequencing) raw data from microalgae studies. MARACAS generates sets of differentially expressed genes or lists of genomic loci for RNA-seq and ChIP-seq analysis respectively. Second, ALGAEFUN (microALGAE FUNctional enrichment tool) is a web-based application where gene sets generated from RNA-seq analysis as well as lists of genomic loci from ChIP-seq analysis can be used as input. On the one hand, it can be used to perform Gene Ontology and biological pathways enrichment analysis over gene sets. On the other hand, using the results of ChIP-seq data analysis, it identifies a set of potential target genes and analyses the distribution of the loci over gene features. Graphical representation of the results as well as tables with gene annotations are generated and can be downloaded for further analysis. CONCLUSIONS: ALGAEFUN with MARACAS provides an integrated environment for the microalgae research community that facilitates the process of obtaining relevant biological information from raw RNA-seq and ChIP-seq data. These applications are designed to assist researchers in the interpretation of gene lists and genomic loci based on functional enrichment analysis. ALGAEFUN with MARACAS is publicly available on https://greennetwork.us.es/AlgaeFUN/ .

BACKGROUND: Identification of biomarkers, which are measurable characteristics of biological datasets, can be challenging. Although amplicon sequence variants (ASVs) can be considered potential biomarkers, identifying important ASVs in high-throughput sequencing datasets is challenging. Noise, algorithmic failures to account for specific distributional properties, and feature interactions can complicate the discovery of ASV biomarkers. In addition, these issues can impact the replicability of various models and elevate false-discovery rates. Contemporary machine learning approaches can be leveraged to address these issues. Ensembles of decision trees are particularly effective at classifying the types of data commonly generated in high-throughput sequencing (HTS) studies due to their robustness when the number of features in the training data is orders of magnitude larger than the number of samples. In addition, when combined with appropriate model introspection algorithms, machine learning algorithms can also be used to discover and select potential biomarkers. However, the construction of these models could introduce various biases which potentially obfuscate feature discovery. RESULTS: We developed a decision tree ensemble, LANDMark, which uses oblique and non-linear cuts at each node. In synthetic and toy tests LANDMark consistently ranked as the best classifier and often outperformed the Random Forest classifier. When trained on the full metabarcoding dataset obtained from Canada's Wood Buffalo National Park, LANDMark was able to create highly predictive models and achieved an overall balanced accuracy score of 0.96 ± 0.06. The use of recursive feature elimination did not impact LANDMark's generalization performance and, when trained on data from the BE amplicon, it was able to outperform the Linear Support Vector Machine, Logistic Regression models, and Stochastic Gradient Descent models (p ≤ 0.05). Finally, LANDMark distinguishes itself due to its ability to learn smoother non-linear decision boundaries. CONCLUSIONS: Our work introduces LANDMark, a meta-classifier which blends the characteristics of several machine learning models into a decision tree and ensemble learning framework. To our knowledge, this is the first study to apply this type of ensemble approach to amplicon sequencing data and we have shown that analyzing these datasets using LANDMark can produce highly predictive and consistent models.

BACKGROUND: Transcriptionally informed predictions are increasingly important for sub-typing cancer patients, understanding underlying biology and to inform novel treatment strategies. For instance, colorectal cancers (CRCs) can be classified into four CRC consensus molecular subgroups (CMS) or five intrinsic (CRIS) sub-types that have prognostic and predictive value. Breast cancer (BRCA) has five PAM50 molecular subgroups with similar value, and the OncotypeDX test provides transcriptomic based clinically actionable treatment-risk stratification. However, assigning samples to these subtypes and other transcriptionally inferred predictions is time consuming and requires significant bioinformatics experience. There is no "universal" method of using data from diverse assay/sequencing platforms to provide subgroup classification using the established classifier sets of genes (CMS, CRIS, PAM50, OncotypeDX), nor one which in provides additional useful functional annotations such as cellular composition, single-sample Gene Set Enrichment Analysis, or prediction of transcription factor activity. RESULTS: To address this bottleneck, we developed classifieR, an easy-to-use R-Shiny based web application that supports flexible rapid single sample annotation of transcriptional profiles derived from cancer patient samples form diverse platforms. We demonstrate the utility of the " classifieR" framework to applications focused on the analysis of transcriptional profiles from colorectal (classifieRc) and breast (classifieRb). Samples are annotated with disease relevant transcriptional subgroups (CMS/CRIS sub-types in classifieRc and PAM50/inferred OncotypeDX in classifieRb), estimation of cellular composition using MCP-counter and xCell, single-sample Gene Set Enrichment Analysis (ssGSEA) and transcription factor activity predictions with Discriminant Regulon Expression Analysis (DoRothEA). CONCLUSIONS: classifieR provides a framework which enables labs without access to a dedicated bioinformation can get information on the molecular makeup of their samples, providing an insight into patient prognosis, druggability and also as a tool for analysis and discovery. Applications are hosted online at https://generatr.qub.ac.uk/app/classifieRc and https://generatr.qub.ac.uk/app/classifieRb after signing up for an account on https://generatr.qub.ac.uk .

BACKGROUND: Delivering tools for genome analysis to users is often difficult given the complex dependencies and conflicts of such tools. Container virtualization systems (such as Singularity) isolate environments, thereby helping developers package tools. However, these systems lack mutual composability, i.e., an easy way to integrate multiple tools in different containers and/or on the host. Another issue is that one may be unable to use a single container system of the same version at all the sites being used, thus discouraging the use of container systems. RESULTS: We developed LPMX, an open-source pure rootless composable container system that provides composability; i.e., the system allows users to easily integrate tools from different containers or even from the host. LPMX accelerates science by letting researchers compose existing containers and containerize tools/pipelines that are difficult to package/containerize using Conda or Singularity, thereby saving researchers' precious time. The technique used in LPMX allows LPMX to run purely in userspace without root privileges even during installation, thus ensuring that we can use LPMX at any Linux clusters with major distributions. The lowest overhead for launching containers with LPMX gives us courage to isolate tools as much as possible into small containers, thereby minimizing the chance of conflicts. The support for the layered file system keeps the total size of container images for a single genomic pipeline modest, as opposed to Singularity, which uses mostly a flat single-layer image. CONCLUSIONS: LPMX is pure rootless container engine with mutual composability, thus saving researchers' time, and accelerating science.

BACKGROUND: In shotgun proteomics, database search engines have been developed to assign peptides to tandem mass (MS/MS) spectra and at the same time post-processing (or rescoring) approaches over the search results have been proposed to increase the number of confident peptide identifications. The most popular post-processing approaches such as Percolator and PeptideProphet have improved rates of peptide identifications by combining multiple scores from database search engines while applying machine learning techniques. Existing post-processing approaches, however, are limited when dealing with results from new search engines because their features for machine learning must be optimized specifically for each search engine. RESULTS: We propose a universal post-processing tool, called TIDD, which supports confident peptide identifications regardless of the search engine adopted. TIDD can work for any (including newly developed) search engines because it calculates universal features that assess peptide-spectrum match quality while it allows additional features provided by search engines (or users) as well. Even though it relies on universal features independent of search tools, TIDD showed similar or better performance than Percolator in terms of peptide identification. TIDD identified 10.23-38.95% more PSMs than target-decoy estimation for MSFragger, which is not supported by Percolator. TIDD offers an easy-to-use simple graphical user interface for user convenience. CONCLUSIONS: TIDD successfully eliminated the requirement for an optimal feature engineering per database search tool, and thus, can be applied directly to any database search results including newly developed ones.

BACKGROUND: RNA sequencing is currently the method of choice for genome-wide profiling of gene expression. A popular approach to quantify expression levels of genes from RNA-seq data is to map reads to a reference genome and then count mapped reads to each gene. Gene annotation data, which include chromosomal coordinates of exons for tens of thousands of genes, are required for this quantification process. There are several major sources of gene annotations that can be used for quantification, such as Ensembl and RefSeq databases. However, there is very little understanding of the effect that the choice of annotation has on the accuracy of gene expression quantification in an RNA-seq analysis. RESULTS: In this paper, we present results from our comparison of Ensembl and RefSeq human annotations on their impact on gene expression quantification using a benchmark RNA-seq dataset generated by the SEQC consortium. We show that the use of RefSeq gene annotation models led to better quantification accuracy, based on the correlation with ground truths including expression data from >800 real-time PCR validated genes, known titration ratios of gene expression and microarray expression data. We also found that the recent expansion of the RefSeq annotation has led to a decrease in its annotation accuracy. Finally, we demonstrated that the RNA-seq quantification differences observed between different annotations were not affected by the use of different normalization methods. CONCLUSION: In conclusion, our study found that the use of the conservative RefSeq gene annotation yields better RNA-seq quantification results than the more comprehensive Ensembl annotation. We also found that, surprisingly, the recent expansion of the RefSeq database, which was primarily driven by the incorporation of sequencing data into the gene annotation process, resulted in a reduction in the accuracy of RNA-seq quantification.

BACKGROUND: Biological sequence clustering is a complicated data clustering problem owing to the high computation costs incurred for pairwise sequence distance calculations through sequence alignments, as well as difficulties in determining parameters for deriving robust clusters. While current approaches are successful in reducing the number of sequence alignments performed, the generated clusters are based on a single sequence identity threshold applied to every cluster. Poor choices of this identity threshold would thus lead to low quality clusters. There is however little support provided to users in selecting thresholds that are well matched with the input sequences. RESULTS: We present a novel sequence clustering approach called ALFATClust that exploits rapid pairwise alignment-free sequence distance calculations and community detection in graph for clusters generation. Instead of a single threshold applied to every generated cluster, ALFATClust is capable of dynamically determining the cut-off threshold for each individual cluster by considering both cluster separation and intra-cluster sequence similarity. Benchmarking analysis shows that ALFATClust generally outperforms existing approaches by simultaneously maintaining cluster robustness and substantial cluster separation for the benchmark datasets. The software also provides an evaluation report for verifying the quality of the non-singleton clusters obtained. CONCLUSIONS: ALFATClust is able to generate sequence clusters having high intra-cluster sequence similarity and substantial separation between clusters without having users to decide precise similarity cut-off thresholds.

BACKGROUND: Transcription factor (TF) binding motifs are identified by high throughput sequencing technologies as means to capture Protein-DNA interactions. These motifs are often represented by consensus sequences in form of position weight matrices (PWMs). With ever-increasing pool of TF binding motifs from multiple sources, redundancy issues are difficult to avoid, especially when every source maintains its own database for collection. One solution can be to cluster biologically relevant or similar PWMs, whether coming from experimental detection or in silico predictions. However, there is a lack of efficient tools to cluster PWMs. Assessing quality of PWM clusters is yet another challenge. Therefore, new methods and tools are required to efficiently cluster PWMs and assess quality of clusters. RESULTS: A new Python package Affinity Based Clustering for Position Weight Matrices (abc4pwm) was developed. It efficiently clustered PWMs from multiple sources with or without using DNA-Binding Domain (DBD) information, generated a representative motif for each cluster, evaluated the clustering quality automatically, and filtered out incorrectly clustered PWMs. Additionally, it was able to update human DBD family database automatically, classified known human TF PWMs to the respective DBD family, and performed TF motif searching and motif discovery by a new ensemble learning approach. CONCLUSION: This work demonstrates applications of abc4pwm in the DNA sequence analysis for various high throughput sequencing data using ~ 1770 human TF PWMs. It recovered known TF motifs at gene promoters based on gene expression profiles (RNA-seq) and identified true TF binding targets for motifs predicted from ChIP-seq experiments. Abc4pwm is a useful tool for TF motif searching, clustering, quality assessment and integration in multiple types of sequence data analysis including RNA-seq, ChIP-seq and ATAC-seq.

BACKGROUND: Barcode-based multiplexing methods can be used to increase throughput and reduce batch effects in large single-cell genomics studies. Despite advantages in flexibility of sample collection and scale, there are additional complications in the data deconvolution steps required to assign each cell to their originating samples. RESULTS: To meet computational needs for efficient sample deconvolution, we developed the tools BarCounter and BarMixer that compute barcode counts and deconvolute mixed single-cell data into sample-specific files, respectively. Together, these tools are implemented as the BarWare pipeline to support demultiplexing from large sequencing projects with many wells of hashed 10x Genomics scRNA-seq data. CONCLUSIONS: BarWare is a modular set of tools linked by shell scripting: BarCounter, a computationally efficient barcode sequence quantification tool implemented in C; and BarMixer, an R package for identification of barcoded populations, merging barcoded data from multiple wells, and quality-control reporting related to scRNA-seq data. These tools and a self-contained implementation of the pipeline are freely available for non-commercial use at https://github.com/AllenInstitute/BarWare-pipeline .

BACKGROUND: A birth-death process of which the births follow a hypoexponential distribution with L phases and are controlled by an on/off mechanism, is a population process which we call the on/off-seq-L process. It is a suitable model for the dynamics of a population of RNA molecules in a single living cell. Motivated by this biological application, our aim is to develop a statistical method to estimate the model parameters of the on/off-seq-L process, based on observations of the population size at discrete time points, and to apply this method to real RNA data. METHODS: It is shown that the on/off-seq-L process can be seen as a quasi birth-death process, and an Erlangization technique can be used to approximate the corresponding likelihood function. An extensive simulation-based numerical study is carried out to investigate the performance of the resulting estimation method. RESULTS AND CONCLUSION: A statistical method is presented to find maximum likelihood estimates of the model parameters for the on/off-seq-L process. Numerical complications related to the likelihood maximization are identified and analyzed, and solutions are presented. The proposed estimation method is a highly accurate method to find the parameter estimates. Based on real RNA data, the on/off-seq-3 process emerges as the best model to describe RNA transcription.

BACKGROUND: Necrotizing enterocolitis (NEC) is a common, potentially catastrophic intestinal disease among very low birthweight premature infants. Affecting up to 15% of neonates born weighing less than 1500 g, NEC causes sudden-onset, progressive intestinal inflammation and necrosis, which can lead to significant bowel loss, multi-organ injury, or death. No unifying cause of NEC has been identified, nor is there any reliable biomarker that indicates an individual patient's risk of the disease. Without a way to predict NEC in advance, the current medical strategy involves close clinical monitoring in an effort to treat babies with NEC as quickly as possible before irrecoverable intestinal damage occurs. In this report, we describe a novel machine learning application for generating dynamic, individualized NEC risk scores based on intestinal microbiota data, which can be determined from sequencing bacterial DNA from otherwise discarded infant stool. A central insight that differentiates our work from past efforts was the recognition that disease prediction from stool microbiota represents a specific subtype of machine learning problem known as multiple instance learning (MIL). RESULTS: We used a neural network-based MIL architecture, which we tested on independent datasets from two cohorts encompassing 3595 stool samples from 261 at-risk infants. Our report also introduces a new concept called the "growing bag" analysis, which applies MIL over time, allowing incorporation of past data into each new risk calculation. This approach allowed early, accurate NEC prediction, with a mean sensitivity of 86% and specificity of 90%. True-positive NEC predictions occurred an average of 8 days before disease onset. We also demonstrate that an attention-gated mechanism incorporated into our MIL algorithm permits interpretation of NEC risk, identifying several bacterial taxa that past work has associated with NEC, and potentially pointing the way toward new hypotheses about NEC pathogenesis. Our system is flexible, accepting microbiota data generated from targeted 16S or "shotgun" whole-genome DNA sequencing. It performs well in the setting of common, potentially confounding preterm neonatal clinical events such as perinatal cardiopulmonary depression, antibiotic administration, feeding disruptions, or transitions between breast feeding and formula. CONCLUSIONS: We have developed and validated a robust MIL-based system for NEC prediction from harmlessly collected premature infant stool. While this system was developed for NEC prediction, our MIL approach may also be applicable to other diseases characterized by changes in the human microbiota.

BACKGROUND: DCLEAR is an R package used for single cell lineage reconstruction. The advances of CRISPR-based gene editing technologies have enabled the prediction of cell lineage trees based on observed edited barcodes from each cell. However, the performance of existing reconstruction methods of cell lineage trees was not accessed until recently. In response to this problem, the Allen Institute hosted the Cell Lineage Reconstruction Dream Challenge in 2020 to crowdsource relevant knowledge from across the world. Our team won sub-challenges 2 and 3 in the challenge competition. RESULTS: The DCLEAR package contained the R codes, which was submitted in response to sub-challenges 2 and 3. Our method consists of two steps: (1) distance matrix estimation and (2) the tree reconstruction from the distance matrix. We proposed two novel methods for distance matrix estimation as outlined in the DCLEAR package. Using our method, we find that two of the more sophisticated distance methods display a substantially improved level of performance compared to the traditional Hamming distance method. DCLEAR is open source and freely available from R CRAN and from under the GNU General Public License, version 3. CONCLUSIONS: DCLEAR is a powerful resource for single cell lineage reconstruction.

BACKGROUND: Named Entity Recognition (NER) and Normalisation (NEN) are core components of any text-mining system for biomedical texts. In a traditional concept-recognition pipeline, these tasks are combined in a serial way, which is inherently prone to error propagation from NER to NEN. We propose a parallel architecture, where both NER and NEN are modeled as a sequence-labeling task, operating directly on the source text. We examine different harmonisation strategies for merging the predictions of the two classifiers into a single output sequence. RESULTS: We test our approach on the recent Version 4 of the CRAFT corpus. In all 20 annotation sets of the concept-annotation task, our system outperforms the pipeline system reported as a baseline in the CRAFT shared task, a competition of the BioNLP Open Shared Tasks 2019. We further refine the systems from the shared task by optimising the harmonisation strategy separately for each annotation set. CONCLUSIONS: Our analysis shows that the strengths of the two classifiers can be combined in a fruitful way. However, prediction harmonisation requires individual calibration on a development set for each annotation set. This allows achieving a good trade-off between established knowledge (training set) and novel information (unseen concepts).

Erratum in BMC Bioinformatics. 2022 Jun 17;23(1):238.

BACKGROUND: lncRNAs play a critical role in numerous biological processes and life activities, especially diseases. Considering that traditional wet experiments for identifying uncovered lncRNA-disease associations is limited in terms of time consumption and labor cost. It is imperative to construct reliable and efficient computational models as addition for practice. Deep learning technologies have been proved to make impressive contributions in many areas, but the feasibility of it in bioinformatics has not been adequately verified. RESULTS: In this paper, a machine learning-based model called LDACE was proposed to predict potential lncRNA-disease associations by combining Extreme Learning Machine (ELM) and Convolutional Neural Network (CNN). Specifically, the representation vectors are constructed by integrating multiple types of biology information including functional similarity and semantic similarity. Then, CNN is applied to mine both local and global features. Finally, ELM is chosen to carry out the prediction task to detect the potential lncRNA-disease associations. The proposed method achieved remarkable Area Under Receiver Operating Characteristic Curve of 0.9086 in Leave-one-out cross-validation and 0.8994 in fivefold cross-validation, respectively. In addition, 2 kinds of case studies based on lung cancer and endometrial cancer indicate the robustness and efficiency of LDACE even in a real environment. CONCLUSIONS: Substantial results demonstrated that the proposed model is expected to be an auxiliary tool to guide and assist biomedical research, and the close integration of deep learning and biology big data will provide life sciences with novel insights.

BACKGROUND: Ontologies of precisely defined, controlled vocabularies are essential to curate the results of biological experiments such that the data are machine searchable, can be computationally analyzed, and are interoperable across the biomedical research continuum. There is also an increasing need for methods to interrelate phenotypic data easily and accurately from experiments in animal models with human development and disease. RESULTS: Here we present the Xenopus phenotype ontology (XPO) to annotate phenotypic data from experiments in Xenopus, one of the major vertebrate model organisms used to study gene function in development and disease. The XPO implements design patterns from the Unified Phenotype Ontology (uPheno), and the principles outlined by the Open Biological and Biomedical Ontologies (OBO Foundry) to maximize interoperability with other species and facilitate ongoing ontology management. Constructed in Web Ontology Language (OWL) the XPO combines the existing uPheno library of ontology design patterns with additional terms from the Xenopus Anatomy Ontology (XAO), the Phenotype and Trait Ontology (PATO) and the Gene Ontology (GO). The integration of these different ontologies into the XPO enables rich phenotypic curation, whilst the uPheno bridging axioms allows phenotypic data from Xenopus experiments to be related to phenotype data from other model organisms and human disease. Moreover, the simple post-composed uPheno design patterns facilitate ongoing XPO development as the generation of new terms and classes of terms can be substantially automated. CONCLUSIONS: The XPO serves as an example of current best practices to help overcome many of the inherent challenges in harmonizing phenotype data between different species. The XPO currently consists of approximately 22,000 terms and is being used to curate phenotypes by Xenbase, the Xenopus Model Organism Knowledgebase, forming a standardized corpus of genotype-phenotype data that can be directly related to other uPheno compliant resources.

BACKGROUND: Given the economic and environmental importance of allopolyploids and other species with highly duplicated genomes, there is a need for methods to distinguish paralogs, i.e. duplicate sequences within a genome, from Mendelian loci, i.e. single copy sequences that pair at meiosis. The ratio of observed to expected heterozygosity is an effective tool for filtering loci but requires genotyping to be performed first at a high computational cost, whereas counting the number of sequence tags detected per genotype is computationally quick but very ineffective in inbred or polyploid populations. Therefore, new methods are needed for filtering paralogs. RESULTS: We introduce a novel statistic, Hind/HE, that uses the probability that two reads sampled from a genotype will belong to different alleles, instead of observed heterozygosity. The expected value of Hind/HE is the same across all loci in a dataset, regardless of read depth or allele frequency. In contrast to methods based on observed heterozygosity, it can be estimated and used for filtering loci prior to genotype calling. In addition to filtering paralogs, it can be used to filter loci with null alleles or high overdispersion, and identify individuals with unexpected ploidy and hybrid status. We demonstrate that the statistic is useful at read depths as low as five to 10, well below the depth needed for accurate genotype calling in polyploid and outcrossing species. CONCLUSIONS: Our methodology for estimating Hind/HE across loci and individuals, as well as determining reasonable thresholds for filtering loci, is implemented in polyRAD v1.6, available at https://github.com/lvclark/polyRAD . In large sequencing datasets, we anticipate that the ability to filter markers and identify problematic individuals prior to genotype calling will save researchers considerable computational time.

BACKGROUND: The prediction of protein secondary structures is a crucial and significant step for ab initio tertiary structure prediction which delivers the information about proteins activity and functions. As the experimental methods are expensive and sometimes impossible, many SS predictors, mainly based on different machine learning methods have been proposed for many years. Currently, most of the top methods use evolutionary-based input features produced by PSSM and HHblits software, although quite recently the embeddings-the new description of protein sequences generated by language models (LM) have appeared that could be leveraged as input features. Apart from input features calculation, the top models usually need extensive computational resources for training and prediction and are barely possible to run on a regular PC. SS prediction as the imbalanced classification problem should not be judged by the commonly used Q3/Q8 metrics. Moreover, as the benchmark datasets are not random samples, the classical statistical null hypothesis testing based on the Neyman-Pearson approach is not appropriate. RESULTS: We present a lightweight deep network ProteinUnet2 for SS prediction which is based on U-Net convolutional architecture and evolutionary-based input features (from PSSM and HHblits) as well as SPOT-Contact features. Through an extensive evaluation study, we report the performance of ProteinUnet2 in comparison with top SS prediction methods based on evolutionary information (SAINT and SPOT-1D). We also propose a new statistical methodology for prediction performance assessment based on the significance from Fisher-Pitman permutation tests accompanied by practical significance measured by Cohen's effect size. CONCLUSIONS: Our results suggest that ProteinUnet2 architecture has much shorter training and inference times while maintaining results similar to SAINT and SPOT-1D predictors. Taking into account the relatively long times of calculating evolutionary-based features (from PSSM in particular), it would be worth conducting the predictive ability tests on embeddings as input features in the future. We strongly believe that our proposed here statistical methodology for the evaluation of SS prediction results will be adopted and used (and even expanded) by the research community.

BACKGROUND: Although both copy number variations (CNVs) and single nucleotide variations (SNVs) detected by single-cell RNA sequencing (scRNA-seq) are used to study intratumor heterogeneity and detect clonal groups, a software that integrates these two types of data in the same cells is unavailable. RESULTS: We developed Clonal Architecture with Integration of SNV and CNV (CAISC), an R package for scRNA-seq data analysis that clusters single cells into distinct subclones by integrating CNV and SNV genotype matrices using an entropy weighted approach. The performance of CAISC was tested on simulation data and four real datasets, which confirmed its high accuracy in sub-clonal identification and assignment, including subclones which cannot be identified using one type of data alone. Furthermore, integration of SNV and CNV allowed for accurate examination of expression changes between subclones, as demonstrated by the results from trisomy 8 clones of the myelodysplastic syndromes (MDS) dataset. CONCLUSIONS: CAISC is a powerful tool for integration of CNV and SNV data from scRNA-seq to identify clonal clusters with better accuracy than obtained from a single type of data. CAISC allows users to interactively examine clonal assignments.

BACKGROUND: Genetic risk scores (GRS) summarize genetic features such as single nucleotide polymorphisms (SNPs) in a single statistic with respect to a given trait. So far, GRS are typically built using generalized linear models or regularized extensions. However, these linear methods are usually not able to incorporate gene-gene interactions or non-linear SNP-response relationships. Tree-based statistical learning methods such as random forests and logic regression may be an alternative to such regularized-regression-based methods and are investigated in this article. Moreover, we consider modifications of random forests and logic regression for the construction of GRS. RESULTS: In an extensive simulation study and an application to a real data set from a German cohort study, we show that both tree-based approaches can outperform elastic net when constructing GRS for binary traits. Especially a modification of logic regression called logic bagging could induce comparatively high predictive power as measured by the area under the curve and the statistical power. Even when considering no epistatic interaction effects but only marginal genetic effects, the regularized regression method lead in most cases to inferior results. CONCLUSIONS: When constructing GRS, we recommend taking random forests and logic bagging into account, in particular, if it can be assumed that possibly unknown epistasis between SNPs is present. To develop the best possible prediction models, extensive joint hyperparameter optimizations should be conducted.

BACKGROUND: Protein-protein interactions have pivotal roles in life processes, and aberrant interactions are associated with various disorders. Interaction site identification is key for understanding disease mechanisms and design new drugs. Effective and efficient computational methods for the PPI prediction are of great value due to the overall cost of experimental methods. Promising results have been obtained using machine learning methods and deep learning techniques, but their effectiveness depends on protein representation and feature selection. RESULTS: We define a new abstraction of the protein structure, called hierarchical representations, considering and quantifying spatial and sequential neighboring among amino acids. We also investigate the effect of molecular abstractions using the Graph Convolutional Networks technique to classify amino acids as interface and no-interface ones. Our study takes into account three abstractions, hierarchical representations, contact map, and the residue sequence, and considers the eight functional classes of proteins extracted from the Protein-Protein Docking Benchmark 5.0. The performance of our method, evaluated using standard metrics, is compared to the ones obtained with some state-of-the-art protein interface predictors. The analysis of the performance values shows that our method outperforms the considered competitors when the considered molecules are structurally similar. CONCLUSIONS: The hierarchical representation can capture the structural properties that promote the interactions and can be used to represent proteins with unknown structures by codifying only their sequential neighboring. Analyzing the results, we conclude that classes should be arranged according to their architectures rather than functions.

BACKGROUND: Third-generation sequencing offers some advantages over next-generation sequencing predecessors, but with the caveat of harboring a much higher error rate. Clustering-related sequences is an essential task in modern biology. To accurately cluster sequences rich in errors, error type and frequency need to be accounted for. Levenshtein distance is a well-established mathematical algorithm for measuring the edit distance between words and can specifically weight insertions, deletions and substitutions. However, there are drawbacks to using Levenshtein distance in a biological context and hence has rarely been used for this purpose. We present novel modifications to the Levenshtein distance algorithm to optimize it for clustering error-rich biological sequencing data. RESULTS: We successfully introduced a bidirectional frameshift allowance with end-user determined accommodation caps combined with weighted error discrimination. Furthermore, our modifications dramatically improved the computational speed of Levenstein distance. For simulated ONT MinION and PacBio Sequel datasets, the average clustering sensitivity for 3GOLD was 41.45% (S.D. 10.39) higher than Sequence-Levenstein distance, 52.14% (S.D. 9.43) higher than Levenshtein distance, 55.93% (S.D. 8.67) higher than Starcode, 42.68% (S.D. 8.09) higher than CD-HIT-EST and 61.49% (S.D. 7.81) higher than DNACLUST. For biological ONT MinION data, 3GOLD clustering sensitivity was 27.99% higher than Sequence-Levenstein distance, 52.76% higher than Levenshtein distance, 56.39% higher than Starcode, 48% higher than CD-HIT-EST and 70.4% higher than DNACLUST. CONCLUSION: Our modifications to Levenshtein distance have improved its speed and accuracy compared to the classic Levenshtein distance, Sequence-Levenshtein distance and other commonly used clustering approaches on simulated and biological third-generation sequenced datasets. Our clustering approach is appropriate for datasets of unknown cluster centroids, such as those generated with unique molecular identifiers as well as known centroids such as barcoded datasets. A strength of our approach is high accuracy in resolving small clusters and mitigating the number of singletons.

BACKGROUND/AIM: In higher eukaryotes, the three-dimensional (3D) organization of the genome is intimately related to numerous key biological functions including gene expression, DNA repair and DNA replication regulations. Alteration of 3D organization, in particular topologically associating domains (TADs), is detrimental to the organism and can give rise to a broad range of diseases such as cancers. METHODS: Here, we propose a versatile regression framework which not only identifies TADs in a fast and accurate manner, but also detects differential TAD borders across conditions for which few methods exist, and predicts 3D genome reorganization after chromosomal rearrangement. Moreover, the framework is biologically meaningful, has an intuitive interpretation and is easy to visualize. RESULT AND CONCLUSION: The novel regression ranks among top TAD callers. Moreover, it identifies new features of the genome we called TAD facilitators, and that are enriched with specific transcription factors. It also unveils the importance of cell-type specific transcription factors in establishing novel TAD borders during neuronal differentiation. Lastly, it compares favorably with the state-of-the-art method for predicting rearranged 3D genome.

BACKGROUND: Cell and circadian cycles control a large fraction of cell and organismal physiology by regulating large periodic transcriptional programs that encompass anywhere from 15 to 80% of the genome despite performing distinct functions. In each case, these large periodic transcriptional programs are controlled by gene regulatory networks (GRNs), and it has been shown through genetics and chromosome mapping approaches in model systems that at the core of these GRNs are small sets of genes that drive the transcript dynamics of the GRNs. However, it is unlikely that we have identified all of these core genes, even in model organisms. Moreover, large periodic transcriptional programs controlling a variety of processes certainly exist in important non-model organisms where genetic approaches to identifying networks are expensive, time-consuming, or intractable. Ideally, the core network components could be identified using data-driven approaches on the transcriptome dynamics data already available. RESULTS: This study shows that a unified set of quantified dynamic features of high-throughput time series gene expression data are more prominent in the core transcriptional regulators of cell and circadian cycles than in their outputs, in multiple organism, even in the presence of external periodic stimuli. Additionally, we observe that the power to discriminate between core and non-core genes is largely insensitive to the particular choice of quantification of these features. CONCLUSIONS: There are practical applications of the approach presented in this study for network inference, since the result is a ranking of genes that is enriched for core regulatory elements driving a periodic phenotype. In this way, the method provides a prioritization of follow-up genetic experiments. Furthermore, these findings reveal something unexpected-that there are shared dynamic features of the transcript abundance of core components of unrelated GRNs that control disparate periodic phenotypes.

BACKGROUND: The accuracy of protein 3D structure prediction has been dramatically improved with the help of advances in deep learning. In the recent CASP14, Deepmind demonstrated that their new version of AlphaFold (AF) produces highly accurate 3D models almost close to experimental structures. The success of AF shows that the multiple sequence alignment of a sequence contains rich evolutionary information, leading to accurate 3D models. Despite the success of AF, only the prediction code is open, and training a similar model requires a vast amount of computational resources. Thus, developing a lighter prediction model is still necessary. RESULTS: In this study, we propose a new protein 3D structure modeling method, A-Prot, using MSA Transformer, one of the state-of-the-art protein language models. An MSA feature tensor and row attention maps are extracted and converted into 2D residue-residue distance and dihedral angle predictions for a given MSA. We demonstrated that A-Prot predicts long-range contacts better than the existing methods. Additionally, we modeled the 3D structures of the free modeling and hard template-based modeling targets of CASP14. The assessment shows that the A-Prot models are more accurate than most top server groups of CASP14. CONCLUSION: These results imply that A-Prot accurately captures the evolutionary and structural information of proteins with relatively low computational cost. Thus, A-Prot can provide a clue for the development of other protein property prediction methods.

BACKGROUND: Upland cotton provides the most natural fiber in the world. During fiber development, the quality and yield of fiber were influenced by gene transcription. Revealing sequence features related to transcription has a profound impact on cotton molecular breeding. We applied convolutional neural networks to predict gene expression status based on the sequences of gene transcription start regions. After that, a gradient-based interpretation and an N-adjusted kernel transformation were implemented to extract sequence features contributing to transcription. RESULTS: Our models had approximate 80% accuracies, and the area under the receiver operating characteristic curve reached over 0.85. Gradient-based interpretation revealed 5' untranslated region contributed to gene transcription. Furthermore, 6 DOF binding motifs and 4 transcription activator binding motifs were obtained by N-adjusted kernel-motif transformation from models in three developmental stages. Apart from 10 general motifs, 3 DOF5.1 genes were also detected. In silico analysis about these motifs' binding proteins implied their potential functions in fiber formation. Besides, we also found some novel motifs in plants as important sequence features for transcription. CONCLUSIONS: In conclusion, the N-adjusted kernel transformation method could interpret convolutional neural networks and reveal important sequence features related to transcription during fiber development. Potential functions of motifs interpreted from convolutional neural networks could be validated by further wet-lab experiments and applied in cotton molecular breeding.

BACKGROUND: Current protein family modeling methods like profile Hidden Markov Model (pHMM), k-mer based methods, and deep learning-based methods do not provide very accurate protein function prediction for proteins in the twilight zone, due to low sequence similarity to reference proteins with known functions. RESULTS: We present a novel method EnsembleFam, aiming at better function prediction for proteins in the twilight zone. EnsembleFam extracts the core characteristics of a protein family using similarity and dissimilarity features calculated from sequence homology relations. EnsembleFam trains three separate Support Vector Machine (SVM) classifiers for each family using these features, and an ensemble prediction is made to classify novel proteins into these families. Extensive experiments are conducted using the Clusters of Orthologous Groups (COG) dataset and G Protein-Coupled Receptor (GPCR) dataset. EnsembleFam not only outperforms state-of-the-art methods on the overall dataset but also provides a much more accurate prediction for twilight zone proteins. CONCLUSIONS: EnsembleFam, a machine learning method to model protein families, can be used to better identify members with very low sequence homology. Using EnsembleFam protein functions can be predicted  using just sequence information with better accuracy than state-of-the-art methods.