Search

Visual individual recognition requires animals to distinguish among conspecifics based on appearance. Though visual individual recognition has been reported in a range of taxa, the features that animals rely on to discriminate between individuals are often not well understood. Northern paper wasp females, Polistes fuscatus, possess individually distinctive color patterns on their faces, which mediate individual recognition. It is currently unclear what facial features P. fuscatus use to distinguish individuals. The anterior optic tubercle, a chromatic processing brain region, is especially sensitive to social experience during development, suggesting that color may be important for recognition in this species. We sought to test the roles of color in wasp facial recognition. Color may be important simply because it creates a pattern. If this is the case, then wasps should perform similarly when discriminating color or grayscale images of the same faces. Alternatively, color itself may be important for recognition, which would predict poorer performance on grayscale image discrimination relative to color images. We found wasps trained on grayscale faces, unlike those trained on color images, did not perform better than chance. Suggesting that color is necessary for the recognition of an image as a face by the wasp visual system.

We present a deep learning-based framework to design and quantify point-of-care sensors. As its proof-of-concept and use-case, we demonstrated a low-cost and rapid paper-based vertical flow assay (VFA) for high sensitivity C-Reactive Protein (hsCRP) testing, a common medical test used for quantifying the degree of inflammation in patients at risk of cardio-vascular disease (CVD). A machine learning-based sensor design framework was developed for two key tasks: (1) to determine an optimal configuration of immunoreaction spots and conditions, spatially-multiplexed on a paper-based sensing membrane, and (2) to accurately infer the target analyte concentration based on the signals of the optimal VFA configuration. Using a custom-designed mobile-phone based VFA reader, a clinical study was performed with 85 human serum samples to characterize the quantification accuracy around the clinically defined cutoffs for CVD risk stratification. Results from blindly-tested VFAs indicate a competitive coefficient of variation of 11.2% with a linearity of R2 = 0.95; in addition to the success in the high-sensitivity CRP range (i.e., 0-10 mg/L), our results further demonstrate a mitigation of the hook-effect at higher CRP concentrations due to the incorporation of antigen capture spots within the multiplexed sensing membrane of the VFA. This paper-based computational VFA that is powered by deep learning could expand access to CVD health screening, and the presented machine learning-enabled sensing framework can be broadly used to design cost-effective and mobile sensors for various point-of-care diagnostics applications.

Background and ObjectiveQuestionnaires are essential tools in many scientific fields, including health and medicine. However, the analysis of paper-and-pencil questionnaires is time consuming, source of errors and expensive, limiting its use in large cohort studies. Computer-based questionnaires might be a valuable alternative but they may introduce bias, especially for sensitive questions, and they require programming skills. The aim of this study is to develop a reliable and adaptable open-source technique (i.e. LightQuest) to automatically analyse various types of scanned paper-and-pencil questionnaires with closed questions, including those with inverted scale.MethodsTo evaluate the usefulness of LightQuest, the time needed for 7 experimenters for manually code 10 sets of 4 frequently used questionnaires and the number of errors (i.e. reliability) were compared with the time and errors their made using LightQuest.ResultsLightQuest was twice as fast as the manual analysis, even though the time to create the reference model was taken into account (933s vs. 1935s, t(2)=8.81, p-1 for the manual technique versus 0.55s.question-1 for LightQuest (t(2)=22.5, pConclusionLightQuest demonstrated clear superiority both in terms of time and reliability. The script of this first open-source technique, which does not require programming skills, is downloadable in supplemental data and may become an asset for all studies using questionnaires.

Good scientific writing is essential to career development and to the progress of science. A well-structured manuscript allows readers and reviewers to get excited about the subject matter, to understand and verify the paper’s contributions, and to integrate these contributions into a broader context. However, many scientists struggle with producing high-quality manuscripts and typically get little training in paper writing. Focusing on how readers consume information, we present a set of 10 simple rules to help you get across the main idea of your paper. These rules are designed to make your paper more influential and the process of writing more efficient and pleasurable.

Motivation With the advancement of technology, we can generate and access large-scale, high dimensional and diverse genomics data, especially through single-cell RNA sequencing (scRNA-seq). However, integrative downstream analysis from multiple scRNA-seq datasets remains challenging due to batch effects. Results In this article, we propose a light-structured deep learning framework called ResPAN for scRNA-seq data integration. ResPAN is based on Wasserstein Generative Adversarial Network (WGAN) combined with random walk mutual nearest neighbor pairing and fully skip-connected autoencoders to reduce the differences among batches. We also discuss the limitations of existing methods and demonstrate the advantages of our model over seven other methods through extensive benchmarking studies on both simulated data under various scenarios and real datasets across different scales. Our model achieves leading performance on both batch correction and biological information conservation and maintains scalable to datasets with over half a million cells. Availability and implementation An open-source implementation of ResPAN and scripts to reproduce the results can be downloaded from: https://github.com/AprilYuge/ResPAN. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Inter-protein (interfacial) contact prediction is very useful for in silico structural characterization of protein–protein interactions. Although deep learning has been applied to this problem, its accuracy is not as good as intra-protein contact prediction. Results We propose a new deep learning method GLINTER (Graph Learning of INTER-protein contacts) for interfacial contact prediction of dimers, leveraging a rotational invariant representation of protein tertiary structures and a pretrained language model of multiple sequence alignments. Tested on the 13th and 14th CASP-CAPRI datasets, the average top L/10 precision achieved by GLINTER is 54% on the homodimers and 52% on all the dimers, much higher than 30% obtained by the latest deep learning method DeepHomo on the homodimers and 15% obtained by BIPSPI on all the dimers. Our experiments show that GLINTER-predicted contacts help improve selection of docking decoys. Availability and implementation The software is available at https://github.com/zw2x/glinter. The datasets are available at https://github.com/zw2x/glinter/data. Supplementary information Supplementary data are available at Bioinformatics online.

Artificial intelligence (AI)‐based image analysis has increased drastically in recent years. However, all applications use individual solutions, highly specialized for a particular task. Here, an easy‐to‐use, adaptable, and open source software, called AIDeveloper (AID) to train neural nets (NN) for image classification without the need for programming is presented. AID provides a variety of NN‐architectures, allowing to apply trained models on new data, obtain performance metrics, and export final models to different formats. AID is benchmarked on large image datasets (CIFAR‐10 and Fashion‐MNIST). Furthermore, models are trained to distinguish areas of differentiated stem cells in images of cell culture. A conventional blood cell count and a blood count obtained using an NN are compared, trained on >1.2 million images, and demonstrated how AID can be used for label‐free classification of B‐ and T‐cells. All models are generated by non‐programmers on generic computers, allowing for an interdisciplinary use., In this work, AIDeveloper (AID), a software for training neural nets for image classification without need for programming is introduced. AID is open source and executables are provided for easy installation. The authors demonstrate AID by training models for classification of blood cells and natural images as well as detection of differentiated stem cells, reaching state of the art performance.

MotivationIn silico drug–target interaction (DTI) prediction is important for drug discovery and drug repurposing. Approaches to predict DTIs can proceed indirectly, top-down, using phenotypic effects of drugs to identify potential drug targets, or they can be direct, bottom-up and use molecular information to directly predict binding potentials. Both approaches can be combined with information about interaction networks.ResultsWe developed DTI-Voodoo as a computational method that combines molecular features and ontology-encoded phenotypic effects of drugs with protein–protein interaction networks, and uses a graph convolutional neural network to predict DTIs. We demonstrate that drug effect features can exploit information in the interaction network whereas molecular features do not. DTI-Voodoo is designed to predict candidate drugs for a given protein; we use this formulation to show that common DTI datasets contain intrinsic biases with major affects on performance evaluation and comparison of DTI prediction methods. Using a modified evaluation scheme, we demonstrate that DTI-Voodoo improves significantly over state of the art DTI prediction methods.AvailabilityDTI-Voodoo source code and data necessary to reproduce results are freely available at https://github.com/THinnerichs/DTI-VOODOO.Contacttilman.hinnerichs@kaust.edu.saSupplementary informationSupplementary data are available at https://github.com/THinnerichs/DTI-VOODOO.

Motivation Fundamental to biological study is identifying regulatory interactions. The recent surge in time-series data collection in biology provides a unique opportunity to infer regulations computationally. However, when components oscillate, model-free inference methods, while easily implemented, struggle to distinguish periodic synchrony and causality. Alternatively, model-based methods test the reproducibility of time series given a specific model but require inefficient simulations and have limited applicability. Results We develop an inference method based on a general model of molecular, neuronal and ecological oscillatory systems that merges the advantages of both model-based and model-free methods, namely accuracy, broad applicability and usability. Our method successfully infers the positive and negative regulations within various oscillatory networks, e.g. the repressilator and a network of cofactors at the pS2 promoter, outperforming popular inference methods. Availability and implementation We provide a computational package, ION (Inferring Oscillatory Networks), that users can easily apply to noisy, oscillatory time series to uncover the mechanisms by which diverse systems generate oscillations. Accompanying MATLAB code under a BSD-style license and examples are available at https://github.com/Mathbiomed/ION. Additionally, the code is available under a CC-BY 4.0 License at https://doi.org/10.6084/m9.figshare.16431408.v1. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Antibodies are one of the most important classes of pharmaceuticals, with over 80 approved molecules currently in use against a wide variety of diseases. The drug discovery process for antibody therapeutic candidates however is time- and cost-intensive and heavily reliant on in vivo and in vitro high throughput screens. Here, we introduce a framework for structure-based deep learning for antibodies (DLAB) which can virtually screen putative binding antibodies against antigen targets of interest. DLAB is built to be able to predict antibody–antigen binding for antigens with no known antibody binders. Results We demonstrate that DLAB can be used both to improve antibody–antigen docking and structure-based virtual screening of antibody drug candidates. DLAB enables improved pose ranking for antibody docking experiments as well as selection of antibody–antigen pairings for which accurate poses are generated and correctly ranked. We also show that DLAB can identify binding antibodies against specific antigens in a case study. Our results demonstrate the promise of deep learning methods for structure-based virtual screening of antibodies. Availability and implementation The DLAB source code and pre-trained models are available at https://github.com/oxpig/dlab-public. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Accurate prediction of residue–residue distances is important for protein structure prediction. We developed several protein distance predictors based on a deep learning distance prediction method and blindly tested them in the 14th Critical Assessment of Protein Structure Prediction (CASP14). The prediction method uses deep residual neural networks with the channel-wise attention mechanism to classify the distance between every two residues into multiple distance intervals. The input features for the deep learning method include co-evolutionary features as well as other sequence-based features derived from multiple sequence alignments (MSAs). Three alignment methods are used with multiple protein sequence/profile databases to generate MSAs for input feature generation. Based on different configurations and training strategies of the deep learning method, five MULTICOM distance predictors were created to participate in the CASP14 experiment. Results Benchmarked on 37 hard CASP14 domains, the best performing MULTICOM predictor is ranked 5th out of 30 automated CASP14 distance prediction servers in terms of precision of top L/5 long-range contact predictions [i.e. classifying distances between two residues into two categories: in contact ( Availability and implementation The software package, source code and data of DeepDist2 are freely available at https://github.com/multicom-toolbox/deepdist and https://zenodo.org/record/4712084#.YIIM13VKhQM. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Gene-gene co-expression networks (GCN) are of biological interest for the useful information they provide for understanding gene-gene interactions. The advent of single cell RNA-sequencing allows us to examine more subtle gene co-expression occurring within a cell type. Many imputation and denoising methods have been developed to deal with the technical challenges observed in single cell data; meanwhile, several simulators have been developed for benchmarking and assessing these methods. Most of these simulators, however, either do not incorporate gene co-expression or generate co-expression in an inconvenient manner. Results Therefore, with the focus on gene co-expression, we propose a new simulator, ESCO, which adopts the idea of the copula to impose gene co-expression, while preserving the highlights of available simulators, which perform well for simulation of gene expression marginally. Using ESCO, we assess the performance of imputation methods on GCN recovery and find that imputation generally helps GCN recovery when the data are not too sparse, and the ensemble imputation method works best among leading methods. In contrast, imputation fails to help in the presence of an excessive fraction of zero counts, where simple data aggregating methods are a better choice. These findings are further verified with mouse and human brain cell data. Availability and implementation The ESCO implementation is available as R package ESCO. Users can either download the development version via github (https://github.com/JINJINT/ESCO) or the archived version via Zenodo (https://zenodo.org/record/4455890). Supplementary information Supplementary data are available at Bioinformatics online.

MotivationConvolutional neural networks (CNNs) have outperformed conventional methods in modeling the sequence specificity of DNA-protein binding. While previous studies have built a connection between CNNs and probabilistic models, simple models of CNNs cannot achieve sufficient accuracy on this problem. Recently, some methods of neural networks have increased performance using complex neural networks whose results cannot be directly interpreted. However, it is difficult to combine probabilistic models and CNNs effectively to improve DNA-protein binding predictions.ResultsIn this paper, we present a novel global pooling method: expectation pooling for predicting DNA-protein binding. Our pooling method stems naturally from the EM algorithm, and its benefits can be interpreted both statistically and via deep learning theory. Through experiments, we demonstrate that our pooling method improves the prediction performance DNA-protein binding. Our interpretable pooling method combines probabilistic ideas with global pooling by taking the expectations of inputs without increasing the number of parameters. We also analyze the hyperparameters in our method and propose optional structures to help fit different datasets. We explore how to effectively utilize these novel pooling methods and show that combining statistical methods with deep learning is highly beneficial, which is promising and meaningful for future studies in this field.Contactdengmh@pku.edu.cn, gaog@mail.cbi.pku.edu.cnSupplementary informationAll code is public in https://github.com/gao-lab/ePooling

Motivation Transferring knowledge between species is challenging: different species contain distinct proteomes and cellular architectures, which cause their proteins to carry out different functions via different interaction networks. Many approaches to protein functional annotation use sequence similarity to transfer knowledge between species. These approaches cannot produce accurate predictions for proteins without homologues of known function, as many functions require cellular context for meaningful prediction. To supply this context, network-based methods use protein-protein interaction (PPI) networks as a source of information for inferring protein function and have demonstrated promising results in function prediction. However, most of these methods are tied to a network for a single species, and many species lack biological networks. Results In this work, we integrate sequence and network information across multiple species by computing IsoRank similarity scores to create a meta-network profile of the proteins of multiple species. We use this integrated multispecies meta-network as input to train a maxout neural network with Gene Ontology terms as target labels. Our multispecies approach takes advantage of more training examples, and consequently leads to significant improvements in function prediction performance compared to two network-based methods, a deep learning sequence-based method and the BLAST annotation method used in the Critial Assessment of Functional Annotation. We are able to demonstrate that our approach performs well even in cases where a species has no network information available: when an organism’s PPI network is left out we can use our multi-species method to make predictions for the left-out organism with good performance. Availability and implementation The code is freely available at https://github.com/nowittynamesleft/NetQuilt. The data, including sequences, PPI networks and GO annotations are available at https://string-db.org/. Supplementary information Supplementary data are available at Bioinformatics online.

Automated segmentation templates can save clinicians time compared to de novo segmentation but may still take substantial time to review and correct. It has not been thoroughly investigated which automated segmentation-corrected segmentation similarity metrics best predict clinician correction time. Bilateral thoracic cavity volumes in 329 CT scans were segmented by a UNet-inspired deep learning segmentation tool and subsequently corrected by a fourth-year medical student. Eight spatial similarity metrics were calculated between the automated and corrected segmentations and associated with correction times using Spearman’s rank correlation coefficients. Nine clinical variables were also associated with metrics and correction times using Spearman’s rank correlation coefficients or Mann–Whitney U tests. The added path length, false negative path length, and surface Dice similarity coefficient correlated better with correction time than traditional metrics, including the popular volumetric Dice similarity coefficient (respectively ρ = 0.69, ρ = 0.65, ρ = − 0.48 versus ρ = − 0.25; correlation p values

Motivation The latest advances in cancer sequencing, and the availability of a wide range of methods to infer the evolutionary history of tumors, have made it important to evaluate, reconcile and cluster different tumor phylogenies. Recently, several notions of distance or similarities have been proposed in the literature, but none of them has emerged as the golden standard. Moreover, none of the known similarity measures is able to manage mutations occurring multiple times in the tree, a circumstance often occurring in real cases. Results To overcome these limitations, in this article, we propose MP3, the first similarity measure for tumor phylogenies able to effectively manage cases where multiple mutations can occur at the same time and mutations can occur multiple times. Moreover, a comparison of MP3 with other measures shows that it is able to classify correctly similar and dissimilar trees, both on simulated and on real data. Availability and implementation An open source implementation of MP3 is publicly available at https://github.com/AlgoLab/mp3treesim. Supplementary information Supplementary data are available at Bioinformatics online.

Small non-coding RNAs such as microRNAs are master regulators of gene expression. One of the most promising applications of miRNAs is the use as liquid biopsy. Especially early diagnosis is an effective means to increase patients’ overall survival. E.g. in oncology a tumor is detected at best prior to its clinical manifestation. We generated genome-wide miRNA profiles from serum of patients and controls from the population-based Janus Serum Bank (JSB) and analyzed them by bioinformatics and artificial intelligence approaches. JSB contains sera from 318,628 originally healthy persons, more than 96,000 of whom later developed cancer. We selected 210 serum samples of patients with lung, colon or breast cancer at three time points prior to diagnosis, after cancer diagnosis and controls. The controls were matched with regard to age of the blood donor and to the time points of blood drawing, which were 27, 32, or 38 years prior to diagnosis. Using ANOVA we report 70 significantly deregulated markers (adjusted p-value−10). Further, 91miRNAs were differently expressed in pre-diagnostic samples as compared to controls (nominal p

A self-supervised workflow uses a 2D class average to progressively train a convolutional neural network for automated particle picking in cryo-EM., High-resolution single-particle cryo-EM data analysis relies on accurate particle picking. To facilitate the particle picking process, a self-supervised workflow has been developed. This includes an iterative strategy, which uses a 2D class average to improve training particles, and a progressively improved convolutional neural network for particle picking. To automate the selection of particles, a threshold is defined (%/Res) using the ratio of percentage class distribution and resolution as a cutoff. This workflow has been tested using six publicly available data sets with different particle sizes and shapes, and can automatically pick particles with minimal user input. The picked particles support high-resolution reconstructions at 3.0 Å or better. This workflow is a step towards automated single-particle cryo-EM data analysis at the stage of particle picking. It may be used in conjunction with commonly used single-particle analysis packages such as Relion, cryoSPARC, cisTEM, SPHIRE and EMAN2.

Motivation One of the major goals in large-scale genomic studies is to identify genes with a prognostic impact on time-to-event outcomes which provide insight into the disease process. With rapid developments in high-throughput genomic technologies in the past two decades, the scientific community is able to monitor the expression levels of tens of thousands of genes and proteins resulting in enormous datasets where the number of genomic features is far greater than the number of subjects. Methods based on univariate Cox regression are often used to select genomic features related to survival outcome; however, the Cox model assumes proportional hazards (PH), which is unlikely to hold for each feature. When applied to genomic features exhibiting some form of non-proportional hazards (NPH), these methods could lead to an under- or over-estimation of the effects. We propose a broad array of marginal screening techniques that aid in feature ranking and selection by accommodating various forms of NPH. First, we develop an approach based on Kullback–Leibler information divergence and the Yang–Prentice model that includes methods for the PH and proportional odds (PO) models as special cases. Next, we propose R2 measures for the PH and PO models that can be interpreted in terms of explained randomness. Lastly, we propose a generalized pseudo-R2 index that includes PH, PO, crossing hazards and crossing odds models as special cases and can be interpreted as the percentage of separability between subjects experiencing the event and not experiencing the event according to feature measurements. Results We evaluate the performance of our measures using extensive simulation studies and publicly available datasets in cancer genomics. We demonstrate that the proposed methods successfully address the issue of NPH in genomic feature selection and outperform existing methods. Availability and implementation R code for the proposed methods is available at github.com/lburns27/Feature-Selection. Contact karthik.devarajan@fccc.edu Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Network diffusion and label propagation are fundamental tools in computational biology, with applications like gene–disease association, protein function prediction and module discovery. More recently, several publications have introduced a permutation analysis after the propagation process, due to concerns that network topology can bias diffusion scores. This opens the question of the statistical properties and the presence of bias of such diffusion processes in each of its applications. In this work, we characterized some common null models behind the permutation analysis and the statistical properties of the diffusion scores. We benchmarked seven diffusion scores on three case studies: synthetic signals on a yeast interactome, simulated differential gene expression on a protein–protein interaction network and prospective gene set prediction on another interaction network. For clarity, all the datasets were based on binary labels, but we also present theoretical results for quantitative labels. Results Diffusion scores starting from binary labels were affected by the label codification and exhibited a problem-dependent topological bias that could be removed by the statistical normalization. Parametric and non-parametric normalization addressed both points by being codification-independent and by equalizing the bias. We identified and quantified two sources of bias—mean value and variance—that yielded performance differences when normalizing the scores. We provided closed formulae for both and showed how the null covariance is related to the spectral properties of the graph. Despite none of the proposed scores systematically outperformed the others, normalization was preferred when the sought positive labels were not aligned with the bias. We conclude that the decision on bias removal should be problem and data-driven, i.e. based on a quantitative analysis of the bias and its relation to the positive entities. Availability The code is publicly available at https://github.com/b2slab/diffuBench and the data underlying this article are available at https://github.com/b2slab/retroData Supplementary information Supplementary data are available at Bioinformatics online.

MotivationFeature selection can improve the accuracy of machine learning models, but appropriate steps must be taken to avoid overfitting. Nested cross-validation (nCV) is a common approach that chooses the classification model and features to represent a given outer fold based on features that give the maximum inner-fold accuracy. Differential privacy is a related technique to avoid overfitting that uses a privacy preserving noise mechanism to identify features that are stable between training and holdout sets.MethodsWe develop consensus nested CV (cnCV) that combines the idea of feature stability from differential privacy with nested CV. Feature selection is applied in each inner fold and the consensus of top features across folds is a used as a measure of feature stability or reliability instead of classification accuracy, which is used in standard nCV. We use simulated data with main effects, correlation, and interactions to compare the classification accuracy and feature selection performance of the new cnCV with standard nCV, Elastic Net optimized by CV, differential privacy, and private Evaporative Cooling (pEC). We also compare these methods using real RNA-Seq data from a study of major depressive disorder.ResultsThe cnCV method has similar training and validation accuracy to nCV, but cnCV has much shorter run times because it does not construct classifiers in the inner folds. The cnCV method chooses a more parsimonious set of features with fewer false positives than nCV. The cnCV method has similar accuracy to pEC and cnCV selects stable features between folds without the need to specify a privacy threshold. We show that cnCV is an effective and efficient approach for combining feature selection with classification.AvailabilityCode available at https://github.com/insilico/cncv.Contactbrett.mckinney@utulsa.eduSupplementary information

Summary Machine learning feature selection methods are needed to detect complex interaction-network effects in complicated modeling scenarios in high-dimensional data, such as GWAS, gene expression, eQTL and structural/functional neuroimage studies for case–control or continuous outcomes. In addition, many machine learning methods have limited ability to address the issues of controlling false discoveries and adjusting for covariates. To address these challenges, we develop a new feature selection technique called Nearest-neighbor Projected-Distance Regression (NPDR) that calculates the importance of each predictor using generalized linear model regression of distances between nearest-neighbor pairs projected onto the predictor dimension. NPDR captures the underlying interaction structure of data using nearest-neighbors in high dimensions, handles both dichotomous and continuous outcomes and predictor data types, statistically corrects for covariates, and permits statistical inference and penalized regression. We use realistic simulations with interactions and other effects to show that NPDR has better precision-recall than standard Relief-based feature selection and random forest importance, with the additional benefit of covariate adjustment and multiple testing correction. Using RNA-Seq data from a study of major depressive disorder (MDD), we show that NPDR with covariate adjustment removes spurious associations due to confounding. We apply NPDR to eQTL data to identify potentially interacting variants that regulate transcripts associated with MDD and demonstrate NPDR’s utility for GWAS and continuous outcomes. Availability and implementation Available at: https://insilico.github.io/npdr/. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Efficient and accurate alignment of DNA/RNA sequence reads to each other or to a reference genome/transcriptome is an important problem in genomic analysis. Nanopore sequencing has emerged as a major sequencing technology and many long-read aligners have been designed for aligning nanopore reads. However, the high error rate makes accurate and efficient alignment difficult. Utilizing the noise and error characteristics inherent in the sequencing process properly can play a vital role in constructing a robust aligner. In this article, we design QAlign, a pre-processor that can be used with any long-read aligner for aligning long reads to a genome/transcriptome or to other long reads. The key idea in QAlign is to convert the nucleotide reads into discretized current levels that capture the error modes of the nanopore sequencer before running it through a sequence aligner. Results We show that QAlign is able to improve alignment rates from around 80% up to 90% with nanopore reads when aligning to the genome. We also show that QAlign improves the average overlap quality by 9.2, 2.5 and 10.8% in three real datasets for read-to-read alignment. Read-to-transcriptome alignment rates are improved from 51.6% to 75.4% and 82.6% to 90% in two real datasets. Availability and implementation https://github.com/joshidhaivat/QAlign.git. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Many biological processes are regulated by single molecules and molecular assemblies within cells that are visible by microscopy as punctate features, often diffraction limited. Here, we present detecting-NEMO (dNEMO), a computational tool optimized for accurate and rapid measurement of fluorescent puncta in fixed-cell and time-lapse images. Results The spot detection algorithm uses the à trous wavelet transform, a computationally inexpensive method that is robust to imaging noise. By combining automated with manual spot curation in the user interface, fluorescent puncta can be carefully selected and measured against their local background to extract high-quality single-cell data. Integrated into the workflow are segmentation and spot-inspection tools that enable almost real-time interaction with images without time consuming pre-processing steps. Although the software is agnostic to the type of puncta imaged, we demonstrate dNEMO using smFISH to measure transcript numbers in single cells in addition to the transient formation of IKK/NEMO puncta from time-lapse images of cells exposed to inflammatory stimuli. We conclude that dNEMO is an ideal user interface for rapid and accurate measurement of fluorescent molecular assemblies in biological imaging data. Availability and implementation The data and software are freely available online at https://github.com/recleelab. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Gene expression and regulation, a key molecular mechanism driving human disease development, remains elusive, especially at early stages. Integrating the increasing amount of population-level genomic data and understanding gene regulatory mechanisms in disease development are still challenging. Machine learning has emerged to solve this, but many machine learning methods were typically limited to building an accurate prediction model as a ‘black box’, barely providing biological and clinical interpretability from the box. Results To address these challenges, we developed an interpretable and scalable machine learning model, ECMarker, to predict gene expression biomarkers for disease phenotypes and simultaneously reveal underlying regulatory mechanisms. Particularly, ECMarker is built on the integration of semi- and discriminative-restricted Boltzmann machines, a neural network model for classification allowing lateral connections at the input gene layer. This interpretable model is scalable without needing any prior feature selection and enables directly modeling and prioritizing genes and revealing potential gene networks (from lateral connections) for the phenotypes. With application to the gene expression data of non-small-cell lung cancer patients, we found that ECMarker not only achieved a relatively high accuracy for predicting cancer stages but also identified the biomarker genes and gene networks implying the regulatory mechanisms in the lung cancer development. In addition, ECMarker demonstrates clinical interpretability as its prioritized biomarker genes can predict survival rates of early lung cancer patients (P-value < 0.005). Finally, we identified a number of drugs currently in clinical use for late stages or other cancers with effects on these early lung cancer biomarkers, suggesting potential novel candidates on early cancer medicine. Availabilityand implementation ECMarker is open source as a general-purpose tool at https://github.com/daifengwanglab/ECMarker. Supplementary information Supplementary data are available at Bioinformatics online.

MotivationThe transcriptomic data is being frequently used in the research of biomarker genes of different diseases and biological states. The most common tasks there are data harmonization and treatment outcome prediction. Both of them can be addressed via the style transfer approach. Either technical factors or any biological details about the samples which we would like to control (gender, biological state, treatment etc.) can be used as style components.ResultsThe proposed style transfer solution is based on Conditional Variational Autoencoders, Y-Autoencoders and adversarial feature decomposition. In order to quantitatively measure the quality of the style transfer, neural network classifiers which predict the style and semantics after training on real expression were used. Comparison with several existing style-transfer based approaches shows that proposed model has the highest style prediction accuracy on all considered datasets while having comparable or the best semantics prediction accuracy.Availabilityhttps://github.com/NRshka/stvae-sourceContactantonec@novel-soft.comSupplementary informationFigShare.com (https://dx.doi.org/10.6084/m9.figshare.9925115)

Motivation Data normalization is an important step in processing proteomics data generated in mass spectrometry experiments, which aims to reduce sample-level variation and facilitate comparisons of samples. Previously published methods for normalization primarily depend on the assumption that the distribution of protein expression is similar across all samples. However, this assumption fails when the protein expression data is generated from heterogenous samples, such as from various tissue types. This led us to develop a novel data-driven method for improved normalization to correct the systematic bias meanwhile maintaining underlying biological heterogeneity. Results To robustly correct the systematic bias, we used the density-power-weight method to down-weigh outliers and extended the one-dimensional robust fitting method described in the previous work to our structured data. We then constructed a robustness criterion and developed a new normalization algorithm, called RobNorm. In simulation studies and analysis of real data from the genotype-tissue expression project, we compared and evaluated the performance of RobNorm against other normalization methods. We found that the RobNorm approach exhibits the greatest reduction in systematic bias while maintaining across-tissue variation, especially for datasets from highly heterogeneous samples. Availabilityand implementation https://github.com/mwgrassgreen/RobNorm. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Transcription factor (TF) binds to the promoter region of a gene to control gene expression. Identifying precise TF binding sites (TFBSs) is essential for understanding the detailed mechanisms of TF-mediated gene regulation. However, there is a shortage of computational approach that can deliver single base pair resolution prediction of TFBS. Results In this paper, we propose DeepSNR, a Deep Learning algorithm for predicting TF binding location at Single Nucleotide Resolution de novo from DNA sequence. DeepSNR adopts a novel deconvolutional network (deconvNet) model and is inspired by the similarity to image segmentation by deconvNet. The proposed deconvNet architecture is constructed on top of ‘DeepBind’ and we trained the entire model using TF-specific data from ChIP-exonuclease (ChIP-exo) experiments. DeepSNR has been shown to outperform motif search–based methods for several evaluation metrics. We have also demonstrated the usefulness of DeepSNR in the regulatory analysis of TFBS as well as in improving the TFBS prediction specificity using ChIP-seq data. Availability and implementation DeepSNR is available open source in the GitHub repository (https://github.com/sirajulsalekin/DeepSNR) Supplementary information Supplementary data are available at Bioinformatics online.

Motivation The precise targeting of antibodies and other protein therapeutics is required for their proper function and the elimination of deleterious off-target effects. Often the molecular structure of a therapeutic target is unknown and randomized methods are used to design antibodies without a model that relates antibody sequence to desired properties. Results Here, we present Ens-Grad, a machine learning method that can design complementarity determining regions of human Immunoglobulin G antibodies with target affinities that are superior to candidates derived from phage display panning experiments. We also demonstrate that machine learning can improve target specificity by the modular composition of models from different experimental campaigns, enabling a new integrative approach to improving target specificity. Our results suggest a new path for the discovery of therapeutic molecules by demonstrating that predictive and differentiable models of antibody binding can be learned from high-throughput experimental data without the need for target structural data. Availability and implementation Sequencing data of the phage panning experiment are deposited at NIH’s Sequence Read Archive (SRA) under the accession number SRP158510. We make our code available at https://github.com/gifford-lab/antibody-2019. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Co-expression of two genes across different conditions is indicative of their involvement in the same biological process. However, when using RNA-Seq datasets with many experimental conditions from diverse sources, only a subset of the experimental conditions is expected to be relevant for finding genes related to a particular Gene Ontology (GO) term. Therefore, we hypothesize that when the purpose is to find similarly functioning genes, the co-expression of genes should not be determined on all samples but only on those samples informative for the GO term of interest. Results To address this, we developed Metric Learning for Co-expression (MLC), a fast algorithm that assigns a GO-term-specific weight to each expression sample. The goal is to obtain a weighted co-expression measure that is more suitable than the unweighted Pearson correlation for applying Guilt-By-Association-based function predictions. More specifically, if two genes are annotated with a given GO term, MLC tries to maximize their weighted co-expression and, in addition, if one of them is not annotated with that term, the weighted co-expression is minimized. Our experiments on publicly available Arabidopsis thaliana RNA-Seq data demonstrate that MLC outperforms standard Pearson correlation in term-centric performance. Moreover, our method is particularly good at more specific terms, which are the most interesting. Finally, by observing the sample weights for a particular GO term, one can identify which experiments are important for learning that term and potentially identify novel conditions that are relevant, as demonstrated by experiments in both A. thaliana and Pseudomonas Aeruginosa. Availability and implementation MLC is available as a Python package at www.github.com/stamakro/MLC. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Automatic biomedical named entity recognition (BioNER) is a key task in biomedical information extraction. For some time, state-of-the-art BioNER has been dominated by machine learning methods, particularly conditional random fields (CRFs), with a recent focus on deep learning. However, recent work has suggested that the high performance of CRFs for BioNER may not generalize to corpora other than the one it was trained on. In our analysis, we find that a popular deep learning-based approach to BioNER, known as bidirectional long short-term memory network-conditional random field (BiLSTM-CRF), is correspondingly poor at generalizing. To address this, we evaluate three modifications of BiLSTM-CRF for BioNER to improve generalization: improved regularization via variational dropout, transfer learning and multi-task learning. Results We measure the effect that each strategy has when training/testing on the same corpus (‘in-corpus’ performance) and when training on one corpus and evaluating on another (‘out-of-corpus’ performance), our measure of the model’s ability to generalize. We found that variational dropout improves out-of-corpus performance by an average of 4.62%, transfer learning by 6.48% and multi-task learning by 8.42%. The maximal increase we identified combines multi-task learning and variational dropout, which boosts out-of-corpus performance by 10.75%. Furthermore, we make available a new open-source tool, called Saber that implements our best BioNER models. Availability and implementation Source code for our biomedical IE tool is available at https://github.com/BaderLab/saber. Corpora and other resources used in this study are available at https://github.com/BaderLab/Towards-reliable-BioNER. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Information extraction by mining the scientific literature is key to uncovering relations between biomedical entities. Most existing approaches based on natural language processing extract relations from single sentence-level co-mentions, ignoring co-occurrence statistics over the whole corpus. Existing approaches counting entity co-occurrences ignore the textual context of each co-occurrence. Results We propose a novel corpus-wide co-occurrence scoring approach to relation extraction that takes the textual context of each co-mention into account. Our method, called CoCoScore, scores the certainty of stating an association for each sentence that co-mentions two entities. CoCoScore is trained using distant supervision based on a gold-standard set of associations between entities of interest. Instead of requiring a manually annotated training corpus, co-mentions are labeled as positives/negatives according to their presence/absence in the gold standard. We show that CoCoScore outperforms previous approaches in identifying human disease–gene and tissue–gene associations as well as in identifying physical and functional protein–protein associations in different species. CoCoScore is a versatile text mining tool to uncover pairwise associations via co-occurrence mining, within and beyond biomedical applications. Availability and implementation CoCoScore is available at: https://github.com/JungeAlexander/cocoscore. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Polygenic risk score (PRS) methods based on genome-wide association studies (GWAS) have a potential for predicting the risk of developing complex diseases and are expected to become more accurate with larger training datasets and innovative statistical methods. The area under the ROC curve (AUC) is often used to evaluate the performance of PRSs, which requires individual genotypic and phenotypic data in an independent GWAS validation dataset. We are motivated to develop methods for approximating AUC of PRSs based on the summary level data of the validation dataset, which will greatly facilitate the development of PRS models for complex diseases. Results We develop statistical methods and an R package SummaryAUC for approximating the AUC and its variance of a PRS when only the summary level data of the validation dataset are available. SummaryAUC can be applied to PRSs with SNPs either genotyped or imputed in the validation dataset. We examined the performance of SummaryAUC using a large-scale GWAS of schizophrenia. SummaryAUC provides accurate approximations to AUCs and their variances. The bias of AUC is typically Availability and implementation https://github.com/lsncibb/SummaryAUC. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation: Drug discovery demands rapid quantification of compound-protein interaction (CPI). However, there is a lack of methods that can predict compound-protein affinity from sequences alone with high applicability, accuracy, and interpretability. Results: We present a seamless integration of domain knowledges and learning-based approaches. Under novel representations of structurally-annotated protein sequences, a semi-supervised deep learning model that unifies recurrent and convolutional neural networks has been proposed to exploit both unlabeled and labeled data, for jointly encoding molecular representations and predicting affinities. Our representations and models outperform conventional options in achieving relative error in IC$_{50}$ within 5-fold for test cases and 20-fold for protein classes not included for training. Performances for new protein classes with few labeled data are further improved by transfer learning. Furthermore, separate and joint attention mechanisms are developed and embedded to our model to add to its interpretability, as illustrated in case studies for predicting and explaining selective drug-target interactions. Lastly, alternative representations using protein sequences or compound graphs and a unified RNN/GCNN-CNN model using graph CNN (GCNN) are also explored to reveal algorithmic challenges ahead. Availability: Data and source codes are available at https://github.com/Shen-Lab/DeepAffinity Supplementary Information: Supplementary data are available at http://shen-lab.github.io/deep-affinity-bioinf18-supp-rev.pdf, https://github.com/Shen-Lab/DeepAffinity

Motivation Low-frequency DNA mutations are often confounded with technical artifacts from sample preparation and sequencing. With unique molecular identifiers (UMIs), most of the sequencing errors can be corrected. However, errors before UMI tagging, such as DNA polymerase errors during end repair and the first PCR cycle, cannot be corrected with single-strand UMIs and impose fundamental limits to UMI-based variant calling. Results We developed smCounter2, a UMI-based variant caller for targeted sequencing data and an upgrade from the current version of smCounter. Compared to smCounter, smCounter2 features lower detection limit that decreases from 1 to 0.5%, better overall accuracy (particularly in non-coding regions), a consistent threshold that can be applied to both deep and shallow sequencing runs, and easier use via a Docker image and code for read pre-processing. We benchmarked smCounter2 against several state-of-the-art UMI-based variant calling methods using multiple datasets and demonstrated smCounter2’s superior performance in detecting somatic variants. At the core of smCounter2 is a statistical test to determine whether the allele frequency of the putative variant is significantly above the background error rate, which was carefully modeled using an independent dataset. The improved accuracy in non-coding regions was mainly achieved using novel repetitive region filters that were specifically designed for UMI data. Availability and implementation The entire pipeline is available at https://github.com/qiaseq/qiaseq-dna under MIT license. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Several tools exist to count Mendelian violations in family trios by comparing variants at the same genomic positions. This naive variant comparison, however, fails to assess regions where multiple variants need to be examined together, resulting in reduced accuracy of existing Mendelian violation checking tools. Results We introduce VBT, a trio concordance analysis tool, which identifies Mendelian violations by approximately solving the 3-way variant matching problem to resolve variant representation differences in family trios. We show that VBT outperforms previous trio comparison methods by accuracy. Availability and implementation VBT is implemented in C++ and source code is available under GNU GPLv3 license at the following URL: https://github.com/sbg/VBT-TrioAnalysis.git. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation The prevalence of high-throughput experimental methods has resulted in an abundance of large-scale molecular and functional interaction networks. The connectivity of these networks provides a rich source of information for inferring functional annotations for genes and proteins. An important challenge has been to develop methods for combining these heterogeneous networks to extract useful protein feature representations for function prediction. Most of the existing approaches for network integration use shallow models that encounter difficulty in capturing complex and highly non-linear network structures. Thus, we propose deepNF, a network fusion method based on Multimodal Deep Autoencoders to extract high-level features of proteins from multiple heterogeneous interaction networks. Results We apply this method to combine STRING networks to construct a common low-dimensional representation containing high-level protein features. We use separate layers for different network types in the early stages of the multimodal autoencoder, later connecting all the layers into a single bottleneck layer from which we extract features to predict protein function. We compare the cross-validation and temporal holdout predictive performance of our method with state-of-the-art methods, including the recently proposed method Mashup. Our results show that our method outperforms previous methods for both human and yeast STRING networks. We also show substantial improvement in the performance of our method in predicting gene ontology terms of varying type and specificity. Availability and implementation deepNF is freely available at: https://github.com/VGligorijevic/deepNF. Supplementary information Supplementary data are available at Bioinformatics online.

The DNA of metazoan somatic cells is packaged into a compact nucleoprotein complex termed chromatin (Van Holde 1988; Wolffe 1998). Chromatin fiber is comprised of highly conserved repetitive units (nucleosomes) that contain an octamer of four core histones and 145–147 base pairs (bp) of DNA wrapped around the octamer in 1.65 turns of a left-handed superhelix (Luger et al. 1997). Nucleosomes are assembled in vivo in an ATP-dependent fashion through a concerted and sequential action of core histone chaperones and motor proteins that belong to the Snf2 family of DNA-dependent ATPases (Gorbalenya and Koonin 1993; Eisen et al. 1995; Mello and Almouzni 2001; Akey and Luger 2003; Haushalter and Kadonaga 2003; Piatti et al. 2011; Torigoe et al. 2013). For instance, Drosophila ACF/CHRAC can mediate chromatin assembly in conjunction with histone chaperone NAP-1 (Varga-Weisz et al. 1997; Ito et al. 1999). Other known ATP-dependent chromatin assembly factors include RSF, CHD1, ATRX, and ToRC/NoRC (LeRoy et al. 1998, 2000; Lusser et al. 2005; Lewis et al. 2010; Emelyanov et al. 2012). The most abundant chromatin component in male germline cells is protamines—small positively charged arginine- and cysteine-rich proteins (Balhorn 2007). During spermiogenesis, protamines replace 85%–95% of DNA-bound histones in the nucleus to achieve a higher density of sperm nuclear DNA (Ward and Coffey 1991; Rathke et al. 2014). Crystalline-like sperm chromatin structure is sixfold more compact than metaphase chromosomes and renders sperm DNA enzymatically inert (Balhorn 1982). At fertilization, the oocyte remodels the condensed sperm chromatin into a transcriptionally competent chromatin of the male pronucleus (McLay and Clarke 2003). During, this process, protamines are expelled and replaced with oocyte-supplied histones, which are then organized into nucleosomes. Sperm chromatin remodeling (SCR) is controlled by biochemical activities in the early oocyte, but components of these activities remain largely unknown (McLay and Clarke 2003). However various protein factors have been implicated in SCR, including core histone chaperones from Xenopus (Nucleoplasmin, N1, HIRA, and TAF-I) and Drosophila (NAP-1, p22, DF31, HIRA, and Yemanuclein) (Dilworth et al. 1987; Philpott and Leno 1992; Kawasaki et al. 1994; Crevel and Cotterill 1995; Ito et al. 1996a; Matsumoto et al. 1999; Loppin et al. 2001; Ray-Gallet et al. 2002; Orsi et al. 2013). In mammals, members of nucleoplasmin/nucleophosmin family proteins (NPM1–3) function in sperm chromatin decondensation in vitro (Okuwaki et al. 2012). In addition, Npm2 knockout female mice exhibit fertility defects consistent with a role of NPM2 in nuclear and nucleolar chromatin organization (Burns et al. 2003). It was also suggested that sperm chromatin decondensation is ATP-dependent (Wright and Longo 1988). Drosophila melanogaster sperm cells contain two major protamines (A and B) encoded by male-specific transcripts Mst35Ba and Mst35Bb, respectively (Ashburner et al. 1999; Jayaramaiah Raja and Renkawitz-Pohl 2005). The Drosophila maternal effect mutant sesame (ssm) prevents male pronucleus formation (Loppin et al. 2001). ssm encodes the histone variant H3.3-specific chaperone HIRA (Tagami et al. 2004), postulated to be required for replication-independent deposition of histones in the male pronucleus during sperm decondensation (Loppin et al. 2005; Bonnefoy et al. 2007). In eggs from homozygous ssm females, maternal histones are not deposited in the chromatin of male pronuclei, preventing normal mitosis and resulting in the development of gynogenetic haploid embryos and embryonic stage lethality. A similar phenotype is observed in null mutants of the gene encoding ATP-dependent chromatin assembly factor CHD1 (Konev et al. 2007). Thus, CHD1 and HIRA act cooperatively and are required for nucleosome assembly during SCR. Intriguingly, protamines are efficiently expelled from the DNA of nascent male pronuclei in Chd1 and ssm eggs (Jayaramaiah Raja and Renkawitz-Pohl 2005; Konev et al. 2007), suggesting that protamine removal and histone deposition are functionally distinct steps. In this study, we used a biochemical approach to identify specific protein components of the Drosophila egg machinery that promote the dissociation of protamine–DNA complexes of sperm chromatin. These factors turn out to be two known core histone chaperones (NAP-1 and NLP), a homolog of mammalian nucleophosmin, and a novel Drosophila histone chaperone (TAP/p32). These putative “protamine chaperones” facilitate SCR independently of CHD1 and HIRA, which mediate nucleosome assembly in nascent male pronuclei. Of note, TAP/p32 is specifically required to expel Drosophila protamine B from sperm chromatin in vitro, whereas NAP-1, NLP, and nucleophosmin share roles in removal of protamine A. We also provide in vivo evidence that TAP/p32 functions in Drosophila egg SCR. In conclusion, we characterized protein factors that mediate the first obligatory step of SCR (protamine dissociation) and reconstituted the complete SCR reaction (reorganization of protamine-containing sperm chromatin into core histone-containing nucleosome arrays) in a purified defined system in vitro.

MotivationThe three-dimensional organization of chromatin plays a critical role in gene regulation and disease. High-throughput chromosome conformation capture experiments such as Hi-C are used to obtain genome-wide maps of 3D chromatin contacts. However, robust estimation of data quality and systematic comparison of these contact maps is challenging due to the multi-scale, hierarchical structure of chromatin contacts and the resulting properties of experimental noise in the data. Measuring concordance of contact maps is important for assessing reproducibility of replicate experiments and for modeling variation between different cellular contexts.ResultsWe introduce a concordance measure called GenomeDISCO (DIfferences between Smoothed COntact maps) for assessing the similarity of a pair of contact maps obtained from chromosome conformation capture experiments. The key idea is to smooth contact maps using random walks on the contact map graph, before estimating concordance. We use simulated datasets to benchmark GenomeDISCO’s sensitivity to different types of noise that affect chromatin contact maps. When applied to a large collection of Hi-C datasets, GenomeDISCO accurately distinguishes biological replicates from samples obtained from different cell types. GenomeDISCO also generalizes to other chromosome conformation capture assays, such as HiChIP.AvailabilitySoftware implementing GenomeDISCO is available at https://github.com/kundajelab/genomedisco.Contactakundaje@stanford.eduSupplementary informationSupplementary data are available at Bioinformatics online.

Recent advances in magnetic sensing has made on-scalp magnetoencephalography (MEG) possible. In particular, optically-pumped magnetometers (OPMs) have reached sensitivity levels that enable their use in MEG. In contrast to the SQUID sensors used in current MEG systems, OPMs do not require cryogenic cooling and can thus be placed within millimetres from the head, enabling the construction of sensor arrays that conform to the shape of an individual’s head. To properly estimate the location of neural sources within the brain, one must accurately know the position and orientation of sensors in relation to the head. With the adaptable on-scalp MEG sensor arrays, this coregistration becomes more challenging than in current SQUID-based MEG systems that use rigid sensor arrays. Here, we used simulations to quantify how accurately one needs to know the position and orientation of sensors in an on-scalp MEG system. The effects that different types of localisation errors have on forward modelling and source estimates obtained by minimum-norm estimation, dipole fitting, and beamforming are detailed. We found that sensor position errors generally have a larger effect than orientation errors and that these errors affect the localisation accuracy of superficial sources the most. To obtain similar or higher accuracy than with current SQUID-based MEG systems, RMS sensor position and orientation errors should be \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$< 4\,\hbox {mm}$$\end{document}

MLST (multi-locus sequence typing) is a classic technique for genotyping bacteria, widely applied for pathogen outbreak surveillance. Traditionally, MLST is based on identifying sequence types from a small number of housekeeping genes. With the increasing availability of whole-genome sequencing (WGS) data, MLST methods have evolved toward larger typing schemes, based on a few hundred genes (core genome MLST, cgMLST) to a few thousand genes (whole genome MLST, wgMLST). Such large-scale MLST schemes have been shown to provide a finer resolution and are increasingly used in various contexts such as hospital outbreaks or foodborne pathogen outbreaks. This methodological shift raises new computational challenges, especially given the large size of the schemes involved. Very few available MLST callers are currently capable of dealing with large MLST schemes.We introduce MentaLiST, a new MLST caller, based on a k-mer voting algorithm and written in the Julia language, specifically designed and implemented to handle large typing schemes. We test it on real and simulated data to show that MentaLiST is faster than any other available MLST caller while providing the same or better accuracy, and is capable of dealing with MLST scheme with up to thousands of genes while requiring limited computational resources. MentaLiST source code and easy installation instructions using a Conda package are available at https://github.com/WGS-TB/MentaLiST.

Motivation Genomic networks represent a complex map of molecular interactions which are descriptive of the biological processes occurring in living cells. Identifying the small over-represented circuitry patterns in these networks helps generate hypotheses about the functional basis of such complex processes. Network motif discovery is a systematic way of achieving this goal. However, a reliable network motif discovery outcome requires generating random background networks which are the result of a uniform and independent graph sampling method. To date, there has been no method to numerically evaluate whether any network motif discovery algorithm performs as intended on realistically sized datasets—thus it was not possible to assess the validity of resulting network motifs. Results In this work, we present IndeCut, the first method to date that characterizes network motif finding algorithm performance in terms of uniform sampling on realistically sized networks. We demonstrate that it is critical to use IndeCut prior to running any network motif finder for two reasons. First, IndeCut indicates the number of samples needed for a tool to produce an outcome that is both reproducible and accurate. Second, IndeCut allows users to choose the tool that generates samples in the most independent fashion for their network of interest among many available options. Availability and implementation The open source software package is available at https://github.com/megrawlab/IndeCut. Supplementary information Supplementary data are available at Bioinformatics online.

Processing of transcripts at the 3’-end involves cleavage at a polyadenylation site followed by the addition of a poly(A)-tail. By selecting which polyadenylation site is cleaved, alternative polyadenylation enables genes to produce transcript isoforms with different 3’-ends. To facilitate the identification and treatment of disease-causing mutations that affect polyadenylation and to understand the underlying regulatory processes, a computational model that can accurately predict polyadenylation patterns based on genomic features is desirable. Previous works have focused on identifying candidate polyadenylation sites and classifying sites which may be tissue-specific. What is lacking is a predictive model of the underlying mechanism of site selection, competition, and processing efficiency in a tissue-specific manner. We develop a deep learning model that trains on 3’-end sequencing data and predicts tissue-specific site selection among competing polyadenylation sites in the 3’ untranslated region of the human genome.Two neural network architectures are evaluated: one built on hand-engineered features, and another that directly learns from the genomic sequence. The hand-engineered features include polyadenylation signals, cis-regulatory elements, n-mer counts, nucleosome occupancy, and RNA-binding protein motifs. The direct-from-sequence model is inferred without prior knowledge on polyadenylation, based on a convolutional neural network trained with genomic sequences surrounding each polyadenylation site as input. Both models are trained using the TensorFlow library.The proposed polyadenylation code can predict site selection among competing polyadenylation sites in different tissues. Importantly, it does so without relying on evolutionary conservation. The model can distinguish pathogenic from benign variants that appear near annotated polyadenylation sites in ClinVar and inspect the genome to find candidate polyadenylation sites. We also provide an analysis on how different features affect the model’s performance.

MotivationTranscription factors (TFs) bind to specific DNA sequence motifs. Several lines of evidence suggest that TF-DNA binding is mediated in part by properties of the local DNA shape: the width of the minor groove, the relative orientations of adjacent base pairs, etc. Several methods have been developed to jointly account for DNA sequence and shape properties in predicting TF binding affinity. However, a limitation of these methods is that they typically require a training set of aligned TF binding sites.ResultsWe describe a sequence+shape kernel that leverages DNA sequence and shape information to better understand protein-DNA binding preference and affinity. This kernel extends an existing class of k-mer based sequence kernels, based on the recently described di-mismatch kernel. Using three in vitro benchmark datasets, derived from universal protein binding microarrays (uPBMs), genomic context PBMs (gcPBMs) and SELEX-seq data, we demonstrate that incorporating DNA shape information improves our ability to predict protein-DNA binding affinity. In particular, we observe that (1) the k-spectrum+shape model performs better than the classical k-spectrum kernel, particularly for small k values; (2) the di-mismatch kernel performs better than the k-mer kernel, for larger k; and (3) the di-mismatch+shape kernel performs better than the di-mismatch kernel for intermediate k values.AvailabilityThe software is available at https://bitbucket.org/wenxiu/sequence-shape.gitContactrohs@usc.edu, william-noble@uw.eduSupplementary informationSupplementary data are available at Bioinformatics online.

Motivation Pseudotime estimation from single-cell gene expression data allows the recovery of temporal information from otherwise static profiles of individual cells. Conventional pseudotime inference methods emphasize an unsupervised transcriptome-wide approach and use retrospective analysis to evaluate the behaviour of individual genes. However, the resulting trajectories can only be understood in terms of abstract geometric structures and not in terms of interpretable models of gene behaviour. Results Here we introduce an orthogonal Bayesian approach termed ‘Ouija’ that learns pseudotimes from a small set of marker genes that might ordinarily be used to retrospectively confirm the accuracy of unsupervised pseudotime algorithms. Crucially, we model these genes in terms of switch-like or transient behaviour along the trajectory, allowing us to understand why the pseudotimes have been inferred and learn informative parameters about the behaviour of each gene. Since each gene is associated with a switch or peak time the genes are effectively ordered along with the cells, allowing each part of the trajectory to be understood in terms of the behaviour of certain genes. We demonstrate that this small panel of marker genes can recover pseudotimes that are consistent with those obtained using the entire transcriptome. Furthermore, we show that our method can detect differences in the regulation timings between two genes and identify ‘metastable’ states—discrete cell types along the continuous trajectories—that recapitulate known cell types. Availability and implementation An open source implementation is available as an R package at http://www.github.com/kieranrcampbell/ouija and as a Python/TensorFlow package at http://www.github.com/kieranrcampbell/ouijaflow. Supplementary information Supplementary data are available at Bioinformatics online.

In cognitive neuroscience, electrical brain activity is most commonly recorded at the scalp. In order to infer the contributions and connectivity of underlying neuronal sources within the brain, it is necessary to reconstruct sensor data at the source level. Several approaches to this reconstruction have been developed, thereby solving the so-called implicit inverse problem Michel et al. (Clin Neurophysiol 115:2195-2222, 2004). However, a unifying premise against which to validate these source reconstructions is seldom available. The dataset provided in this work, in which brain activity is simultaneously recorded on the scalp (non-invasively) by electroencephalography (EEG) and on the cortex (invasively) by electrocorticography (ECoG), can be of a great help in this direction. These multimodal recordings were obtained from a macaque monkey under wakefulness and sedation. Our primary goal was to establish the connectivity architecture between two sources of interest (frontal and parietal), and to assess how their coupling changes over the conditions. We chose these sources because previous studies have shown that the connections between them are modified by anaesthesia Boly et al. (J Neurosci 32:7082-7090, 2012). Our secondary goal was to evaluate the consistency of the connectivity results when analyzing sources recorded from invasive data (128 implanted ECoG sources) and source activity reconstructed from scalp recordings (19 EEG sensors) at the same locations as the ECoG sources. We conclude that the directed connectivity in the frequency domain between cortical sources reconstructed from scalp EEG is qualitatively similar to the connectivity inferred directly from cortical recordings, using both data-driven (directed transfer function) and biologically grounded (dynamic causal modelling) methods. Furthermore, the connectivity changes identified were consistent with previous findings Boly et al. (J Neurosci 32:7082-7090, 2012). Our findings suggest that inferences about directed connectivity based upon non-invasive electrophysiological data have construct validity in relation to invasive recordings.

Motivation Antimicrobial peptides (AMPs) are increasingly being used in the development of new therapeutic drugs in areas such as cancer therapy and hypertension. Additionally, they are seen as an alternative to antibiotics due to the increasing occurrence of bacterial resistance. Wet-laboratory experimental identification, however, is both time-consuming and costly, so in silico models are now commonly used in order to screen new AMP candidates. Results This paper proposes a novel approach for creating model inputs; using pre-trained language models to produce contextualized embeddings, representing the amino acids within each peptide sequence, before a convolutional neural network is trained as the classifier. The results were validated on two datasets—one previously used in AMP prediction research, and a larger independent dataset created by this paper. Predictive accuracies of 93.33% and 88.26% were achieved, respectively, outperforming previous state-of-the-art classification models. Availability and implementation All codes are available and can be accessed here: https://github.com/williamdee1/LMPred_AMP_Prediction. Supplementary information Supplementary data are available at Bioinformatics Advances online.

The modelling of many real-world problems relies on computationally heavy simulations. Since statistical inference rests on repeated simulations to sample the parameter space, the high computational expense of these simulations can become a stumbling block. In this paper, we compare two ways to mitigate this issue based on machine learning methods. One approach is to construct lightweight surrogate models to substitute the simulations used in inference. Alternatively, one might altogether circumnavigate the need for Bayesian sampling schemes and directly estimate the posterior distribution. We focus on stochastic simulations that track autonomous agents and present two case studies of real-world applications: tumour growths and the spread of infectious diseases. We demonstrate that good accuracy in inference can be achieved with a relatively small number of simulations, making our machine learning approaches orders of magnitude faster than classical simulation-based methods that rely on sampling the parameter space. However, we find that while some methods generally produce more robust results than others, no algorithm offers a one-size-fits-all solution when attempting to infer model parameters from observations. Instead, one must choose the inference technique with the specific real-world application in mind. The stochastic nature of the considered real-world phenomena poses an additional challenge that can become insurmountable for some approaches. Overall, we find machine learning approaches that create direct inference machines to be promising for real-world applications. We present our findings as general guidelines for modelling practitioners. Author summaryComputer simulations play a vital role in modern science as they are commonly used to compare theory with observations. One can thus infer the properties of a observed system by comparing the data to the predicted behaviour in different scenarios. Each of these scenarios corresponds to a simulation with slightly different settings. However, since real-world problems are highly complex, the simulations often require extensive computational resources, making direct comparisons with data challenging, if not insurmountable. It is, therefore, necessary to resort to inference methods that mitigate this issue, but it is not clear-cut what path to choose for any specific research problem. In this paper, we provide general guidelines for how to make this choice. We do so by studying examples from oncology and epidemiology and by taking advantage of developments in machine learning. More specifically, we focus on simulations that track the behaviour of autonomous agents, such as single cells or individuals. We show that the best way forward is problem-dependent and highlight the methods that yield the most robust results across the different case studies. We demonstrate that these methods are highly promising and produce reliable results in a small fraction of the time required by classic approaches that rely on comparisons between data and individual simulations. Rather than relying on a single inference technique, we recommend employing several methods and selecting the most reliable based on predetermined criteria.

Choroid plexuses (CP) are structures of the brain ventricles which produce most of the cerebrospinal fluid (CSF). Several postmortem and in vivo studies have pointed towards their role in the inflammatory processes in multiple sclerosis (MS). Automatic segmentation of CP from MRI thus has high value for studying their characteristics in large cohorts of patients. To the best of our knowledge, the only freely available tool for CP segmentation is FreeSurfer but its accuracy for this specific structure is poor. In this paper, we propose to automatically segment CP from non-contrast enhanced T1-weighted MRI. To that end, we introduce a new model called “Axial-MLP” based on an assembly of Axial multi-layer perceptrons (MLPs). This is inspired by recent works which showed that the self-attention layers of Transformers can be replaced with MLPs. This approach is systematically compared with a standard 3D U-Net, nnU-Net, Freesurfer and FastSurfer. For our experiments, we make use of a dataset of 141 subjects (44 controls and 97 patients with MS). We show that all the tested deep learning (DL) methods outperform FreeSurfer (Dice around 0.7 for DL vs 0.33 for FreeSurfer). Axial-MLP is competitive with U-Nets even though it is slightly less accurate. The conclusions of our paper are two-fold: 1) the studied deep learning methods could be useful tools to study CP in large cohorts of MS patients; 2) Axial-MLP is a potentially viable alternative to convolutional neural networks for such tasks, although it could benefit from further improvements.

The COVID-19 pandemic has infected over 200 million people worldwide and killed more than 4 million as of August 2021. Many intervention strategies have been utilized by governments around the world, including masks, social distancing, and vaccinations. However, officials making decisions regarding interventions may have a limited time to act. Computer simulations can aid them by predicting future disease outcomes, but they also have limitations due to requirements on processing power or time. This paper examines whether a machine learning model can be trained on a small subset of simulation runs to inexpensively predict future disease trajectories very close to the original simulation results. Using four previously published agent-based models for COVID-19, this paper analyzes the predictions of decision tree regression machine learning models and compares them to the results of the original simulations. The results indicate that accurate machine learning meta-models can be generated from simulation models with no strong interventions (e.g., vaccines, lockdowns) using small amounts of simulation data. However, meta-models for simulation models that include strong interventions required much more training data to achieve a similar accuracy. This indicates that machine learning meta-models could be used in some scenarios to assist in faster decision making.