Centre for Bioinformatics

AbstractMotivationGenome-scale metabolic network reconstructions (GENREs) are valuable for understanding cellular metabolism in silico. Several tools exist for automatic GENRE generation. However, these tools frequently (1) do not readily integrate with some of the widely-used suites of packaged methods available for network analysis, (2) lack effective network curation tools, (3) are not sufficiently user-friendly, and (4) often produce low quality draft reconstructions.ResultsHere, we present Reconstructor, a user-friendly, COBRApy-compatible tool that produces high quality draft reconstructions with reaction and metabolite naming conventions that are consistent with the ModelSEED biochemistry database and includes a gap-filling technique based on the principles of parsimony. Reconstructor can generate SBML GENREs from three input types: annotated protein .fasta sequences (Type 1 input), a BLASTp output (Type 2), or an existing SBML GENRE that can be further gap-filled (Type 3). While Reconstructor can be used to create GENREs of any species, we demonstrate the utility of Reconstructor with bacterial reconstructions. We demonstrate how Reconstructor readily generates high-quality GENRES that capture strain, species, and higher taxonomic differences in functional metabolism of bacteria and are useful for further biological discovery.AvailabilityThe Reconstructor Python package is freely available for download. Complete installation and usage instructions and benchmarking data are available at http://github.com/emmamglass/reconstructor.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationReplicability is the cornerstone of scientific research. The current statistical method for high-dimensional replicability analysis either cannot control the false discovery rate (FDR) or is too conservative.ResultsWe propose a statistical method, JUMP, for the high-dimensional replicability analysis of two studies. The input is a high-dimensional paired sequence of p-values from two studies and the test statistic is the maximum of p-values of the pair. JUMP uses four states of the p-value pairs to indicate whether they are null or non-null. Conditional on the hidden states, JUMP computes the cumulative distribution function of the maximum of p-values for each state to conservatively approximate the probability of rejection under the composite null of replicability. JUMP estimates unknown parameters and uses a step-up procedure to control FDR. By incorporating different states of composite null, JUMP achieves a substantial power gain over existing methods while controlling the FDR. Analyzing two pairs of spatially resolved transcriptomic datasets, JUMP makes biological discoveries that otherwise cannot be obtained by using existing methods.AvailabilityAn R package JUMP implementing the JUMP method is available on CRAN (https://CRAN.R-project.org/package=JUMP).Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractSummaryAdaptive Immune Receptor Repertoire Sequencing is a rapidly developing field that has advanced understanding of the role of the adaptive immune system in health and disease. Numerous tools have been developed to analyse the complex data produced by this technique but work to compare their accuracy and reliability has been limited. Thorough, systematic assessment of their performance is dependent on the ability to produce high quality simulated datasets with known ground truth. We have developed AIRRSHIP, a flexible and fast Python package that produces synthetic human B cell receptor sequences. AIRRSHIP uses a comprehensive set of reference data to replicate key mechanisms in the immunoglobulin recombination process, with a particular focus on junctional complexity. Repertoires generated by AIRRSHIP are highly similar to published data and all steps in the sequence generation process are recorded. These data can be used to not only determine the accuracy of repertoire analysis tools but can also, by tuning of the large number of user-controllable parameters, give insight into factors that contribute to inaccuracies in results.Availability and ImplementationAIRRSHIP is implemented in Python. It is available via https://github.com/Cowanlab/airrship and on PyPI at https://pypi.org/project/airrship/. Documentation can be found at https://airrship.readthedocs.io/.Supplementary InformationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationTranscriptional profiles of diverse tissues provide significant insights in both fundamental and translational researches, while transcriptome information is not always available for tissues that require invasive biopsies. Alternatively, predicting tissue expression profiles from more accessible ‘surrogate’ samples, especially blood transcriptome, has become a promising strategy when invasive procedures are not practical. However, existing approaches ignore tissue-shared intrinsic relevance, inevitably limiting predictive performance.ResultsWe propose a unified deep learning-based multi-task learning framework, Multi-tissue Transcriptome Mapping (MTM), enabling the prediction of individualized expression profiles from any available tissue of an individual. By jointly leveraging individualized cross-tissue information from reference samples through multi-task learning, MTM achieves superior sample-level and gene-level performance on unseen individuals. With the high prediction accuracy and the ability to preserve individualized biological variations, MTM could facilitate both fundamental and clinical biomedical research.AvailabilityMTM’s code and documentation are available upon publication on GitHub (https://github.com/yangence/MTM).Supplementary informationSupplementary dataSupplementary data will be available at Bioinformatics online.

AbstractMotivationImaging Spatial Transcriptomics (iST) techniques characterize gene expression in cells in their native context by imaging barcoded probes for mRNA with single molecule resolution. However, the need to acquire many rounds of high-magnification imaging data limits the throughput and impact of existing methods.ResultsWe describe the Joint Sparse method for Imaging Transcriptomics (JSIT), an algorithm for decoding lower magnification iST data than that used in standard experimental workflows. JSIT incorporates codebook knowledge and sparsity assumptions into an optimization problem which is less reliant on well separated optical signals than current pipelines. Using experimental data obtained by performing Multiplexed Error-Robust Fluorescence in situ Hybridization (MERFISH) on tissue from mouse brain, we demonstrate that JSIT enables improved throughput and recovery performance over standard decoding methods.AvailabilitySoftware implementation of JSIT, together with example files, are available at https://github.com/jpbryan13/JSIT.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationLong-read transcriptome sequencing (LRTS) has the potential to enhance our understanding of alternative splicing, and the complexity of this process requires the use of versatile computational tools, with the ability to accommodate various stages of the workflow with maximum flexibility.ResultsWe introduce IsoTools, a Python-based LRTS analysis framework that offers a wide range of functionality for transcriptome reconstruction and quantification of transcripts. Furthermore, we integrate a graph-based method for identifying alternative splicing events and a statistical approach based on the beta-binomial distribution for detecting differential events. To demonstrate the effectiveness of our methods, we applied IsoTools to PacBio LRTS data of human hepatocytes treated with the HDAC inhibitor valproic acid. Our results indicate that LRTS can provide valuable insights into alternative splicing, particularly in terms of complex and differential splicing patterns, in comparison to short-read RNA-seq.Availability and ImplementationIsoTools is available on GitHub and PyPI, and its documentation, including tutorials, CLI and API references, can be found at https://isotools.readthedocs.io/.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationSingle-cell RNA sequencing (scRNA-seq) data, annotated by cell type, is useful in a variety of downstream biological applications, such as profiling gene expression at the single-cell level. However, manually assigning these annotations with known marker genes is both time-consuming and subjective.ResultsWe present a Graph Convolutional Network (GCN) based approach to automate the annotation process. Our process builds upon existing labeling approaches, using state-of-the-art tools to find cells with highly confident label assignments through consensus and spreading these confident labels with a semi-supervised GCN. Using simulated data and two scRNA-seq data sets from different tissues, we show that our method improves accuracy over a simple consensus algorithm and the average of the underlying tools. We also compare our method to a non-parametric neighbor majority approach, showing comparable results. We then demonstrate that our GCN method allows for feature interpretation, identifying important genes for cell type classification. We present our completed pipeline, written in PyTorch, as an end-to-end tool for automating and interpreting the classification of scRNA-seq data.AvailabilityOur code for conducting the experiments in this paper and using our model is available at https://github.com/lewinsohndp/scSHARPSupplementary informationSupplementary dataSupplementary data are available at Journal Name online.

AbstractSummarympwR is an R package for a standardized comparison of mass spectrometry (MS)-based proteomic label-free workflows recorded by data-dependent or data-independent spectral acquisition. The user-friendly design allows easy access to compare the influence of sample preparation procedures, combinations of liquid chromatography (LC)-MS setups, as well as intra- and inter-software differences on critical performance measures across an unlimited number of analyses. mpwR supports outputs of commonly used software for bottom-up proteomics, such as ProteomeDiscoverer, Spectronaut, MaxQuant and DIA-NN.AvailabilitympwR is available as an open-source R package. Release versions can be accessed on CRAN (https://CRAN.R-project.org/package=mpwR) for all major operating systems. The development version is maintained on GitHub (https://github.com/okdll/mpwR) and full documentation with examples and workflow templates is provided via the package website (https://okdll.github.io/mpwR/).

AbstractMotivationAn imperative step in drug discovery is the prediction of drug-disease associations (DDAs), which tries to uncover potential therapeutic possibilities for already validated drugs. It is costly and time-consuming to predict DDAs using wet experiments. Graph Neural Networks (GNNs) as an emerging technique have shown superior capacity of dealing with DDA prediction. However, existing GNNs-based DDA prediction methods suffer from sparse supervised signals. As Graph Contrastive Learning (GCL) has shined in mitigating sparse supervised signals, we seek to leverage GCL to enhance the prediction of DDAs. Unfortunately, most conventional GCL-based models corrupt the raw data graph to augment data, which are unsuitable for DDA prediction. Meanwhile, these methods could not model the interactions between nodes effectively, thereby reducing the accuracy of association predictions.ResultsA model is proposed to tap potential drug candidates for diseases, which is called Similarity Measures based Graph Co-contrastive Learning (SMGCL). For learning embeddings from complicated network topologies, SMGCL includes three essential processes: (i) Constructs three views based on similarities between drugs and diseases and DDA information; (ii) Two graph encoders are performed over the three views, so as to model both local and global topologies simultaneously; (iii) A graph co-contrastive learning method is introduced, which co-trains the representations of nodes to maximize the agreement between them, thus generating high-quality prediction results. Contrastive learning serves as an auxiliary task for improving DDA predictions. Evaluated by cross-validations, SMGCL achieves pleasing comprehensive performances. Further proof of the SMGCL’s practicality is provided by case study of Alzheimer’s disease.Availability and Implementationhttps://github.com/Jcmorz/SMGCLSupplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationDrug combination therapy shows significant advantages over monotherapy in cancer treatment. Since the combinational space is difficult to be traversed experimentally, identifying novel synergistic drug combinations based on computational methods has become a powerful tool for pre-screening. Among them, methods based on deep learning have far outperformed other methods. However, most deep learning-based methods are unstable and will give inconsistent predictions even by simply changing the input order of drugs. In addition, the insufficient experimental data of drug combination screening limits the generalization ability of existing models. These problems prevent the deep learning-based models from being in service.ResultsIn this paper, we propose CGMS to address the above problems. CGMS models a drug combination and a cell line as a heterogeneous complete graph, and generates the whole-graph embedding to characterize their interaction by leveraging the heterogeneous graph attention network. Based on the whole-graph embedding, CGMS can make a stable, order-independent prediction. To enhance the generalization ability of CGMS, we apply the multi-task learning technique to train the model on drug synergy prediction task and drug sensitivity prediction task simultaneously. We compare CGMS’s generalization ability with 6 state-of-the-art methods on a public dataset, and CGMS significantly outperforms other methods in the leave-drug combination-out scenario, as well as in the leave-cell line-out and leave-drug-out scenarios. We further present the benefit of eliminating the order dependency and the discrimination power of whole-graph embeddings, interpret the rationality of the attention mechanism, and verify the contribution of multi-task learning.AvailabilityThe code of CGMS is available via https://github.com/TOJSSE-iData/CGMSSupplementary informationSupplementary dataSupplementary data are available at Bioinformatics

AbstractSummaryBiomedical named entity recognition (NER) plays a crucial role in extracting information from documents in biomedical applications. However, many of these applications require NER models to operate at a document level, rather than just a sentence level. This presents a challenge, as the extension from a sentence model to a document model is not always straightforward. Despite the existence of document NER models that are able to make consistent predictions, they still fall short of meeting the expectations of researchers and practitioners in the field. To address this issue, we have undertaken an investigation into the underlying causes of inconsistent predictions. Our research has led us to believe that the use of adjectives and prepositions within entities may be contributing to low label consistency. In this paper, we present our method, ConNER, to enhance a label consistency of modifiers such as adjectives and prepositions. By refining the labels of these modifiers, ConNER is able to improve representations of biomedical entities. The effectiveness of our method is demonstrated on four popular biomedical NER datasets. On three datasets, we achieve a higher F1 score than the previous state-of-the-art model. Our method shows its efficacy on two datasets, resulting in 7.5-8.6% absolute improvements in the F1 score. Our findings suggest that our ConNER method is effective on datasets with intrinsically low label consistency. Through qualitative analysis, we demonstrate how our approach helps the NER model generate more consistent predictions.Availability and implementationOur code and resources are available at https://github.com/dmis-lab/ConNER/.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractSummaryMultimodal single-cell sequencing data provide detailed views into the molecular biology of cells. To allow for interactive analyses of such rich data and to readily derive insights from it, new analysis solutions are required. In this work we present Cellenium, our new scalable visual analytics web application which enables users to semantically integrate and organize all their single-cell RNA-, ATAC-, and CITE-sequencing studies. Users can then find relevant studies and analyze single-cell data within and across studies. An interactive cell annotation feature allows for adding user-defined cell types.Availability and implementationSource code and documentation are freely available under an MIT license and are available on GitHub (https://github.com/Bayer-Group/cellenium). The server backend is implemented in PostgreSQL, Python 3 and GraphQL, the frontend is written in ReactJS, TypeScript and Mantine css, plots are generated using plotlyjs, seaborn, vega-lite and nivo.rocks. The application is dockerized and can be deployed and orchestrated on a standard workstation via docker-compose.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationCancer is a molecular complex and heterogeneous disease. Each type of cancer is usually composed of several subtypes with different treatment responses and clinical outcomes. Therefore, subtyping is a crucial step in cancer diagnosis and therapy. The rapid advances in high-throughput sequencing technologies provide an increasing amount of multi-omics data, which benefits our understanding of cancer genetic architecture, and yet poses new challenges in multi-omics data integration.ResultsWe propose a graph convolutional network model, called MRGCN for multi-omics data integrative representation. MRGCN simultaneously encodes and reconstructs multiple omics expression and similarity relationships into a shared latent embedding space. In addition, MRGCN adopts an indicator matrix to denote the situation of missing values in partial omics, so that the full and partial multi-omics processing procedures are combined in a unified framework. Experimental results on eleven multi-omics datasets show that cancer subtypes obtained by MRGCN with superior enriched clinical parameters and log-rank test P-values in survival analysis over many typical integrative methods.Availabilityhttps://github.com/Polytech-bioinf/MRGCN.git  https://figshare.com/articles/software/MRGCN/23058503Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationThe prediction of reliable Drug–Target Interactions (DTI) is a key task in computer-aided drug design and repurposing. Here we present a new approach based on data fusion for DTI prediction built on top of the NXTfusion library, which generalizes the Matrix Factorization (MF) paradigm by extending it to the nonlinear inference over Entity–Relation (ER) graphs.ResultsWe benchmarked our approach on five data sets and we compared our models against state-of-the-art methods. Our models outperform most of the existing methods and, simultaneously, retain the flexibility to predict both DTIs as binary classification and regression of the real-valued drug–target affinity, competing with models built explicitly for each task. Moreover, our findings suggest that the validation of DTI methods should be stricter than what has been proposed in some previous studies, focusing more on mimicking real-life DTI settings where predictions for previously unseen drugs, proteins and drug–protein pairs are needed. These settings are exactly the context in which the benefit of integrating heterogeneous information with our ER data fusion approach is the most evident.AvailabilityAll software and data is available at https://github.com/eugeniomazzone/CPI-NXTFusion and https://pypi.org/project/NXTfusion/.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationElementary flux modes are a well-known tool for analyzing metabolic networks. The whole set of EFMs cannot be computed in most genome-scale networks due to their large cardinality. Therefore, different methods have been proposed to compute a smaller subset of EFMs that can be used for studying the structure of the network. These latter methods pose the problem of studying the representativeness of the calculated subset. In this paper we present a methodology to tackle this problem.ResultsWe have introduced the concept of stability for a particular network parameter and its relation to the representativeness of the EFM extraction method studied. We have also defined several metrics to study and compare the EFM biases. We have applied these techniques to compare the relative behavior of previously proposed methods in two case studies. Furthermore, we have presented a new method for the EFM computation (PiEFM), which is more stable (less biased) than previous ones, has suitable representativeness measures, and exhibits better variability in the extracted EFMs.Availability and implementationSoftware and additional material is freely available at https://github.com/biogacop/PiEFM

AbstractMotivationLarge-scale prediction of drug-target affinity (DTA) plays an important role in drug discovery. In recent years, machine learning algorithms have made great progress in DTA prediction by utilizing sequence or structural information of both drugs and proteins. However, sequence-based algorithms ignore the structural information of molecules and proteins, while graph-based algorithms are insufficient in feature extraction and information interaction.ResultsIn this paper, we propose NHGNN-DTA, a node-adaptive hybrid neural network for interpretable DTA prediction. It can adaptively acquire feature representations of drugs and proteins and allow information to interact at the graph level, effectively combining the advantages of both sequence-based and graph-based approaches. Experimental results have shown that NHGNN-DTA achieved new state-of-the-art performance. It achieved the mean squared error (MSE) of 0.196 on the Davis dataset (below 0.2 for the first time) and 0.124 on the KIBA dataset (3% improvement). Meanwhile, in the case of cold start scenario, NHGNN-DTA proved to be more robust and more effective with unseen inputs than baseline methods. Furthermore, the multi-head self-attention mechanism endows the model with interpretability, providing new exploratory insights for drug discovery. The case study on Omicron variants of SARS-CoV-2 illustrates the efficient utilization of drug repurposing in COVID-19.AvailabilityThe source code and data are available at https://github.com/hehh77/NHGNN-DTA.

AbstractMotivationTandem mass spectrometry is an essential technology for characterizing chemical compounds at high sensitivity and throughput, and is commonly adopted in many fields. However, computational methods for automated compound identification from their MS/MS spectra are still limited, especially for novel compounds that have not been previously characterized. In recent years, in silico methods were proposed to predict the MS/MS spectra of compounds, which can then be used to expand the reference spectral libraries for compound identification. However, these methods did not consider the compounds’ three-dimensional (3D) conformations, and thus neglected critical structural information.ResultsWe present the 3D Molecular Network for Mass Spectra Prediction (3DMolMS), a deep neural network model to predict the MS/MS spectra of compounds from their 3D conformations. We evaluated the model on the experimental spectra collected in several spectral libraries. The results showed that 3DMolMS predicted the spectra with the average cosine similarity of 0.691 and 0.478 with the experimental MS/MS spectra acquired in positive and negative ion modes, respectively. Furthermore, 3DMolMS model can be generalized to the prediction of MS/MS spectra acquired by different labs on different instruments through minor fine-tuning on a small set of spectra. Finally, we demonstrate that the molecular representation learned by 3DMolMS from MS/MS spectra prediction can be adapted to enhance the prediction of chemical properties such as the elution time (ET) in the liquid chromatography and the collisional cross section (CCS) measured by ion mobility spectrometry (IMS), both of which are often used to improve compound identification.AvailabilityThe codes of 3DMolMS are available at https://github.com/JosieHong/3DMolMS and the web service is at https://spectrumprediction.gnps2.org.

AbstractMotivationNanopore sequencing is emerging as a key pillar in the genomic technology landscape but computational constraints limiting its scalability remain to be overcome. The translation of raw current signal data into DNA or RNA sequence reads, known as ‘basecalling’, is a major friction in any nanopore sequencing workflow. Here, we exploit the advantages of the recently developed signal data format ‘SLOW5’ to streamline and accelerate nanopore basecalling on high-performance computing (HPC) and cloud environments.ResultsSLOW5 permits highly efficient sequential data access, eliminating a potential analysis bottleneck. To take advantage of this, we introduce Buttery-eel, an open-source wrapper for Oxford Nanopore’s Guppy basecaller that enables SLOW5 data access, resulting in performance improvements that are essential for scalable, affordable basecalling.AvailabilityButtery-eel is available at https://github.com/Psy-Fer/buttery-eelSupplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationThe generation of parameter files for molecular dynamics (MD) simulations of small molecules that are suitable for force fields commonly applied to proteins and nucleic acids is often challenging. The ACPYPE software and website aid the generation of such parameter files.ResultsACPYPE uses OpenBabel and ANTECHAMBER to generate MD input files in Gromacs, AMBER, CHARMM and CNS formats. It can now take a SMILES string as input, in addition to the original PDB or mol2 coordinate files, with GAFF2 support and GLYCAM force field conversion added. It can be installed locally via Anaconda, PyPI and Docker distributions, while the web server at https://bio2byte.be/acpype/ was updated with an API, and provides visualization of results for uploaded molecules as well as a pre-generated set of 3738 drug molecules.AvailabilityThe web application is freely available at https://www.bio2byte.be/acpype/ and the open-source code can be found at https://github.com/alanwilter/acpype.Supplementary informationSupplementary dataSupplementary data are available at Bioinformatics online.

AbstractMotivationBone marrow examination is one of the most important indicators in diagnosing hematologic disorders and is typically performed under the microscope via oil-immersion objective lens with a total 100X objective magnification. On the other hand, mitotic detection and identification is critical not only for accurate cancer diagnosis and grading but also for predicting therapy success and survival. Fully automated bone marrow examination and mitotic figure examination from whole-slide images is highly demanded but challenging and poorly explored. Firstly, the complexity and poor reproducibility of microscopic image examination are due to the cell type diversity, delicate intra-lineage discrepancy within the multi-type cell maturation process, cells overlapping, lipid interference and stain variation. Secondly, manual annotation on whole-slide images is tedious, laborious and subject to intra-observer variability, which causes the supervised information restricted to limited, easily identifiable and scattered cells annotated by humans. Thirdly, when the training data is sparsely labeled, many unlabeled objects of interest are wrongly defined as background, which severely confuses AI learners.ResultsThis paper presents an efficient and fully automatic CW-Net approach to address the three issues mentioned above and demonstrates its superior performance on both bone marrow examination and mitotic figure examination. The experimental results demonstrate the robustness and generalizability of the proposed CW-Net on a large bone marrow WSI dataset with 16,456 annotated cells of 19 bone marrow cell types and a large-scale WSI dataset for mitotic figure assessment with 262,481 annotated cells of five cell types.Availability and ImpelmentationAn online web-based system of the proposed method has been created for demonstration. (see https://youtu.be/MRMR25Mls1A).Supplementary InformationSupplementary dataSupplementary data are available at Bioinformatics online.