Computational biology and bioinformatics

Research aimed at improving healthcare has largely focused on male animals and cells. Here, the authors use data from the International Mouse Phenotyping Consortium to show that body weight does not account for all phenotypic differences between male and female mice, supporting more female-focused research.

Previous observational studies of the diabetes drugs metformin vs. sulfonylureas have yielded mixed results about whether metformin reduces the risk of dementia, relative to the sulfonylureas. Here, the authors apply a novel competing risks approach to emulate dementia-related target trials in electronic health records of diabetic patients and a complementary systems pharmacology evaluation on human neural cells.

Here the authors explore the distributional differences expected from distinct biophysical models of transcription and show how measurements from single-cell genomics experiments can shed light on the underlying biological processes.

The mechanism by which DNA methylation might affect complex traits is not well understood. Here, the authors use Mendelian randomization to reveal a substantial role of transcript levels in mediating DNA methylation effects on complex traits and diseases.

Multiregion sequencing is needed to better capture the heterogeneity of hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA). Here, the authors analyse HCC and iCCA tumours with multiregion single-cell RNA-seq, revealing cellular dynamics and communication networks with immune cells.

Software tools for larger-scale intact glycopeptide quantification lag far behind, which hinders exploring the differential sitespecific glycosylation. Here, the authors report pGlycoQuant, a generic tool with a deep learning model for quantitative glycoproteomics at intact glycopeptide level.

Off-target binding hinders the development of therapeutic antibodies and reproducibility in basic research settings. Here the authors develop a method to quantify and reduce the polyreactivity of antibody fragments based on protein sequence alone.

Object detection using machine learning universally requires vast amounts of training datasets. Midtvedt et al. proposes a deep-learning method that enables detecting microscopic objects with sub-pixel accuracy from a single unlabeled image by exploiting the roto-translational symmetries of the problem.

Methods for jointly analysing the different spatial data modalities in 3D are lacking. Here the authors report the computational framework STACI (Spatial Transcriptomic data using over-parameterized graph-based Autoencoders with Chromatin Imaging data) which they apply to an Alzheimer’s disease mouse model.

Nucleosome profiling from cell-free DNA (cfDNA) represents a potential approach for cancer detection and classification. Here, the authors develop Griffin, a computational framework for tumour subtype classification based on cfDNA nucleosome profiling that can work with ultra-low pass sequencing data.

Predicting topological structures from Hi-C data provides insight into comprehending gene expression and regulation. Here, the authors present RefHiC, an attention-based deep learning framework that leverages a reference panel of Hi-C datasets to assist topological structure annotation from a given study sample.

G protein coupled receptors (GPCRs) can couple to different Gα protein subfamilies either selectively or promiscuously. Here, the authors use computational approach to show that selectivity determinants are at the periphery of the GPCR—G protein interface and that promiscuous GPCRs more frequently sample the common rather than selective contacts.

Metabolic labeling is often used to measure protein turnover. Here the authors show that for interconvertible protein species like phosphoforms metabolic labeling does not provide information on turnover differences, but that the relative order of modification can determine the observed dynamics.

The presented Mean-Shift Super Resolution (MSSR) algorithm can extend spatial resolution within a single microscopy image. Its applicability extends across a wide range of experimental and instrumental configurations and it is compatible with other super-resolution microscopy approaches.

Integrating heterogeneous single-cell multi-omics as well as spatially resolved transcriptomic data remains a major challenge. Here the authors report a unified single-cell data integration framework using an unbalanced optimal transport-based deep network.

Identifying the designers of engineered biological sequences would help promote biotechnological innovation while holding designers accountable. Here the authors present the winners of a 2020 data-science competition which improved on previous attempts to attribute plasmid sequences.

Methods that analyse heterogeneity and compare tissue microenvironments using spatial omics data are challenging to develop. Here, the authors present SOTIP, a method that can perform spatial heterogeneity, spatial domain, and differential microenvironment analyses across multiple spatial omics modalities.

Sinonasal tumour diagnosis can be complicated by the heterogeneity of disease and classification systems. Here, the authors use machine learning to classify sinonasal undifferentiated carcinomas into 4 molecular classe with differences in differentiation state and clinical outcome.

Studying the cell composition of acral melanoma at the single-cell level could provide some clues about its poor response to immunotherapy. Here, the authors analyse acral and cutaneous melanoma patient samples using single-cell RNA-sequencing, and reveal a severe immunosuppressive state in acral melanomas

Modelling how endogenous mutations accumulate in tissues is valuable to understand how cancers develop and evolve. Here, the authors establish a mathematical model that can predict the number of endogenous somatic mutations in the lifetime of tissues and approximate the time to cancer development.

Deep learning (DL) has been frequently used in mass spectrometry-based proteomics but there is still a lot of potential. Here, the authors develop a framework that enables building DL models to predict arbitrary peptide properties with only a few lines of code.

The biological underpinnings underlying the increased mortality and morbidity in adolescents and young adults (AYA) remains poorly understood. Here, the authors investigate the clinical and genomic disparities in AYA and older adults in a cohort of more than 100,000 cancer patients.

Division of labour, where members of a group specialise on different tasks, is a central feature of many social organisms. Using a theoretical model, the authors demonstrate that division of labour can emerge spontaneously within a group of entirely identical individuals.

Spatial transcriptomics analyses can be affected by noise and spatial correlation across tissue locations. Here, the authors develop SpatialPCA, a spatially-aware dimensionality reduction method that explicitly models spatial correlation structures, and show its application to the analysis of healthy and tumour tissues.

Mutations in BRCA1/2 are associated with a homologous recombination deficiency phenotype in BRCA-associated cancers. Reversion mutations can restore BRCA1/2 function and result in treatment resistance in these cancer-types. Here, the authors show that, in select cases, reversion mutations in BRCA1/2 can indicate prior BRCA-mediated tumorigenesis in non-canonical histologies.

In this work the authors provide a computational workflow for the parallel, from scratch, design of proteins to rapidly explore the shape diversity of protein folds.

Recovering dropout-affected gene expression values is a challenging problem in bioinformatics. Here, the authors propose a data-driven framework, that first learns the underlying data distribution and then recovers the expression values by imposing a self-consistency on the expression matrix.

The rational design of PROTACs is difficult due to their obscure structure-activity relationship. Here the authors present a deep neural network model - DeepPROTACs - for predicting the degradation capacity of a proposed PROTAC molecule.

Modeling the dynamics of large proteins reveals a fundamental scaling problem. Here, the authors tackle this challenge by decomposing a large system into smaller independent subsystems, simultaneously modeling each subsystem’s kinetics and ensuring their mutual independence.

Gap formation in the vasculature underpins immune and tumour cell infiltration. Here the authors propose a chemo-mechanical model to analyse how feedback between mechanosensitive signalling, active cellular forces and adhesion governs the breakdown, recovery, and integrity of endothelial junctions.

Transcriptome-wide association studies can uncover genes involved in disease. Here, the authors extend the framework with a transcriptome-wide association study approach which incorporates transcription factor occupancy, adding tissue-specific mechanistic support to associations.

Here the authors show that transposable element-mediated rearrangements impact more than 500 kbp of an average human genome, are a source of individual variation, a substrate for evolutionary change, and can occur through diverse mechanisms.

By comprehensively mapping the impact that different classes of mutations (substitutions, insertions, deletions) have on the ability of the amyloid beta peptide to nucleate amyloids, the authors identify a large set of likely pathogenic variants of amyloid beta that are specifically enriched at its polar N-terminal region.

Agonists selectively targeting GPR119 hold promise for treating metabolic disorders. Here, authors reveal that GPR119 adopts a non-canonical consensus structural scaffold with an extended ligand-binding pocket for chemically different agonists.

Development of multiple myeloma is preceded by precursor conditions. Here, the authors use single cell RNA-sequencing of plasma cells from patients across disease stages to identify genomic signatures present even at the earliest stages of disease.

Evolutionary principles could help distinguish driver from passenger mutations in cancer. Here, the authors develop SEISMIC, a method to identify cancer driver genes based on their deviation from expected mutation status patterns across a cohort under neutral evolution, and find potential drivers in melanoma and other cancer types.

Multiple sequence alignments are widely used to predict protein structure, function, and phylogeny, but are uncertain with more diverged sequences. Muscle5 generates ensembles of alternative high-accurate alignments, enabling novel confidence estimates in alignments, trees, and other inferences.

Predicting inter-chain residue-residue distances of protein complexes is useful for constructing and evaluating quaternary structures of the protein complexes. Here, the authors develop a deep attention-based residual network method (CDPred) to predict inter-chain residue-residue distances of protein dimers.

Here, the authors describe two opabiniid-like euarthropods with anterior proboscises from the Middle Ordovician Castle Bank Biota, Wales, UK. Phylogenetic analysis suggests that these specimens may be sister to radiodonts and deuteropods.

Spatially resolved transcriptomics is a relatively new technique that maps transcriptional information within a tissue. Here the authors present MIST, which detects molecular regions from spatially resolved transcriptomics and denoises the missing gene expression values by region-specific imputation.

Current treatment guidelines for Type-2 diabetes endorse a massive number of potential anti-hyper-glycemic treatment options in various permutations and combinations. Here, the authors present a causal deep learning approach for more personalized recommendations of treatment selection.

Comparisons among experimental results with large amounts of data can be more precise and meaningful when done across multiple different conditions simultaneously. Koch et al. introduce a method, called CLIMB, that does this, and captures interpretable and biologically meaningful information.

Seasonal influenza vaccination is an important strategy to prevent serious disease in the elderly, but individual responsiveness to vaccination widely vary. Here authors establish, with an array of state-of-the art methods, the major immunological parameters that distinguish vaccine recipients developing robust antibody response and non-responders

As the throughput of single-cell RNA-seq studies increases, there is a need for tools that can make the data analysis steps more streamlined and convenient. Here, the authors develop UniverSC, a tool that unifies single-cell RNA-seq analysis workflows and also facilitates their use for non-experts.

The organisation of mammalian genomes plays a role in many biological processes. Here the authors report dcHiC, a tool which uses a multivariate distance measure to identify changes in compartmentalisation among multiple genome-wide chromatin contact maps, and apply this to different human and mouse datasets.

Computational modeling and design of G Protein-Coupled Receptor quaternary structures reveals a signaling bias switch at the receptor dimer interface that selectively controls G protein vs β-arrestin activation.

Contaminant sequences in metagenomic samples can potentially impact the interpretation of findings reported in microbiome studies, especially in low biomass environments. Here the authors describe Squeegee, a computational approach designed to detect microbial contamination within low microbial biomass microbiomes and identify microbial contaminants in publicly available datasets that lack negative controls.

Krach et al. dissect the molecular mechanism of the alternative splicing modulator Branaplam in Huntington’s disease. They show that the drug lowers mutant HTT protein levels and ameliorates alternative splicing pathology in an iPSC disease model.

Modification of transcribed mRNAs enables regulation of transcription but its extent in cancer cells is incompletely understood. Here, the authors analyse transcript assembly in over 1000 cancer cell lines and find unannotated transcripts are common, and are associated with drug sensitivity.