Latest bioRxiv papers

Category: bioinformatics — Showing 50 items

• Inferring a novel insecticide resistance metric and exposurevariability in mosquito bioassays across Africa

  Denz, A.; Kont, M. D.; Sanou, A.; Churcher, T. S.; Lambert, B.

  AI Summary

  • This study introduces a new predictive model to assess insecticide resistance in mosquitoes by incorporating data from intensity-dose susceptibility bioassays, addressing variability due to genetic factors.
  • The model was fitted to data from across Africa, focusing on Burkina Faso, to estimate location-specific resistance heterogeneity and exposure differences in bioassays versus experimental huts.
  • The approach aims to enhance malaria transmission models by providing a mechanistic understanding of insecticide resistance's public health impact.

  Abstract

  Malaria claims approximately 500,000 lives each year, and insecticide-treated nets (ITNs), which kill mosquitoes that transmit the disease, remain the most effective intervention. However, resistance to pyrethroids, the primary insecticide class used in ITNs, has risen dramatically in Africa, making it difficult to assess the current public health impact of pyrethroid-ITNs. Past work has modelled the relation between pyrethroid susceptibility measured in discriminating-dose susceptibility bioassays and ITN effectiveness in experimental hut trials. Here, we introduce a new predictive approach that accounts for heterogeneity in insecticide resistance within wild mosquito populations, for example, due to genetic variability, by incorporating data from newly recommended intensity-dose susceptibility bioassays. We fit our mathematical model to a comprehensive data set that combines discriminating dose bioassays from all over Africa, intensity dose bioassays from Burkina Faso, and concurrent experimental hut trials. Our analysis estimates location- and insecticide-specific variation in resistance heterogeneity in Burkina Faso and quantifies differences in insecticide exposure in bioassays and experimental huts. By providing a mechanistic understanding of these experimental data, our approach could be integrated into malaria transmission models to account for the public health impact of insecticide resistance detected by surveillance programmes.

  bioinformatics2026-02-12v4

• GeneReL: A Large Language Model-Powered Platform for Gene Regulatory Relationship Extraction with Community Curation

  Park, J.-S.; Ha, S.; Lee, Y.; Kang, Y. J.

  AI Summary

  • GeneReL is a platform developed to extract and curate gene regulatory relationships in Arabidopsis thaliana using large language models (LLMs) and community validation.
  • It uses a tiered pipeline with different LLMs for screening, extraction, and verification, and includes a five-step gene normalization process.
  • The platform has curated 13,710 interactions, with 86.8% unique compared to IntAct, and features interactive visualization and community voting for validation.

  Abstract

  Motivation: Gene regulatory networks provide fundamental insights into plant biology, yet extracting structured interaction data from scientific literature remains a significant bottleneck. Traditional manual curation cannot scale to meet the demands of modern research, while automated text mining approaches struggle with the complexity of gene nomenclature and relationship classification. Large language models offer promising capabilities for information extraction, but integrated platforms combining LLM extraction with community validation for plant regulatory databases remain scarce. Results: We developed GeneReL, an integrated platform combining LLM-based extraction with community-driven curation for gene regulatory networks in Arabidopsis thaliana. The system employs a tiered pipeline using Claude Haiku 4.5 for screening, Claude Sonnet 4 for extraction, and Claude Opus 4 for verification, along with a novel five-step gene normalization pipeline incorporating paper-text search and LLM-based disambiguation with UniProt annotations. The database contains 13,710 curated interactions across 51 relationship types, with 90.2% classified as high confidence based on linguistic certainty markers in source text. Comparison with IntAct reveals 86.8% of interactions are unique to our literature-derived database, demonstrating complementary coverage to existing resources. The web platform provides card-based browsing with voting capabilities, interactive network visualization using Cytoscape.js with locus-ID-based node consolidation, and administrative interfaces for curator review of ambiguous gene mappings.

  bioinformatics2026-02-12v2

• LineageSim: A Single-Cell Lineage Simulator with Fate-Aware Gene Expression

  Lai, H.; Sadria, M.

  AI Summary

  • LineageSim is introduced as a simulator that generates single-cell lineage data with fate-aware gene expression, addressing the limitation of existing simulators which lack long-range temporal dependencies.
  • The simulator includes latent signals in progenitor states that predict future cell fates, providing a benchmark for cell fate prediction algorithms.
  • Validation through logistic regression showed a 68.3% balanced accuracy, confirming the presence of predictive fate information in the simulated data.

  Abstract

  Single-cell lineage data paired with gene expression are critical for developing computational methods in developmental biology. Since experimental lineage tracing is often technically limited, robust simulations are necessary to provide the ground truth for rigorous validation. However, existing simulators generate largely Markovian gene expression, failing to encode the fate bias observed in real biological systems, where progenitor states exhibit early signatures of future commitment. Consequently, they cannot support the training and evaluation of computational methods that model long-range temporal dependencies. We present LineageSim, a generative framework that introduces fate-aware gene expression, where progenitor states carry latent signals of their descendants' terminal fates. This framework establishes a new class of benchmarks for cell fate prediction algorithms. We validate the presence of these temporal signals by training a logistic regression baseline, which achieves 68.3% balanced accuracy. This confirms that the generated data contain subtle but recoverable fate information, in contrast to existing simulators, where such predictive signals are systematically absent.

  bioinformatics2026-02-12v2

• Reading TEA leaves for de novo protein design

  Pantolini, L.; Durairaj, J.

  AI Summary

  • The study explores de novo protein design using a 20-letter structure-inspired alphabet from protein language model embeddings to enhance Monte Carlo sampling efficiency.
  • This approach allows for rapid template-guided and unconditional design of protein sequences that meet in silico designability criteria, without relying on known homologues.
  • The method significantly reduces the time required for protein design, opening new avenues for therapeutic and industrial applications.

  Abstract

  De novo protein design expands the functional protein universe beyond natural evolution, offering vast therapeutic and industrial potential. Monte Carlo sampling in protein design is under-explored due to the typically long simulation times required or prohibitive time requirements of current structure prediction oracles. Here we make use of a 20-letter structure-inspired alphabet derived from protein language model embeddings to score random mutagenesis-based Metropolis sampling of amino acid sequences. This facilitates fast template-guided and unconditional design, generating sequences that satisfy in silico designability criteria without known homologues. Ultimately, this unlocks a new path to fast and de novo protein design.

  bioinformatics2026-02-12v2

• Predicting interaction-specific protein-protein interaction perturbations by missense variants with MutPred-PPI

  Stewart, R.; Laval, F.; Coppin, G.; Spirohn-Fitzgerald, K.; Tixhon, M.; Hao, T.; Calderwood, M. A.; Mort, M.; Cooper, D. N.; Vidal, M.; Radivojac, P.

  AI Summary

  • MutPred-PPI, a graph attention network, was developed to predict the interaction-specific effects of missense variants on protein-protein interactions using AlphaFold 3-based contact graphs and protein language model embeddings.
  • The model outperformed existing methods with AUCs of 0.85 for seen proteins and 0.72 for unseen proteins, showing strong generalizability.
  • Application to various datasets revealed distinct PPI perturbation patterns, with disease-associated variants showing enrichment for edgetic effects, particularly in cancer and neurodevelopmental disorders.

  Abstract

  Disruption of protein-protein interactions (PPIs) is a major mechanism of a variant's deleterious effect. Computational tools are needed to assess such variants at scale, yet existing predictors rarely consider loss of specific interactions, particularly when variants perturb binding interfaces without significantly affecting protein stability. To address this problem, we present MutPred-PPI, a graph attention network that predicts interaction-specific (edgetic) effects of missense variants by operating on AlphaFold 3-based protein complex contact graphs with protein language model embeddings imposed upon nodes. We systematically evaluated our model with stringent group cross-validation as well as benchmark data recently collected within the IGVF Consortium. MutPred-PPI outperformed all baseline methods across all evaluation criteria, achieving an AUC of 0.85 on seen proteins and 0.72 on previously unseen proteins in cross-validation, demonstrating strong generalizability despite scarce training data. To demonstrate biomedical relevance, we applied MutPred-PPI to variants from ClinVar, HGMD, COSMIC, gnomAD, and two de novo neurodevelopmental disorder-linked datasets. Disease-associated variants from ClinVar and HGMD showed strong enrichment for both quasi-null and edgetic effects, whereas population variants from gnomAD increasingly preserved interactions with higher allele frequencies. Notably, we observed a strong edgetic disruption signature in highly recurrent cancer variants from both the full COSMIC dataset and a subset of variants from oncogenes. Recurrent tumor suppressor gene variants and autism spectrum disorder-associated variants exhibited moderate quasi-null enrichment, whilst neurodevelopmental disorder-linked variants showed a weak edgetic disruption signature. These results indicate distinct PPI perturbation mechanisms across disease types and show that MutPred-PPI captures functionally relevant molecular effects of pathogenic variants.

  bioinformatics2026-02-12v2

• Investigating Enzyme Function by Geometric Matching of Catalytic Motifs

  Hackett, R. E.; Riziotis, I. G.; Larralde, M.; Ribeiro, A. J. M.; Zeller, G.; Thornton, J.

  AI Summary

  • Developed a method using geometric matching to detect catalytic features in protein structures, utilizing a library of 6780 3D coordinate sets from 762 enzyme mechanisms.
  • The approach was validated on 3751 high-quality experimental enzyme structures and predicted human proteome structures, showing higher sensitivity in identifying enzyme homology than sequence or 3D-structure-based methods.
  • This method identifies structural similarities in catalytic sites of divergent enzymes, offering insights into enzyme function evolution, and is available as the Python module Enzyme Motif Miner.

  Abstract

  The rapidly growing universe of predicted protein structures offers opportunities for data driven exploration but requires computationally scalable and interpretable tools. We developed a method to detect catalytic features in protein structures, providing insights into enzyme function and mechanism. A library of 6780 3D coordinate sets describing enzyme catalytic sites, referred to as templates, has been collected from manually curated examples of 762 enzyme catalytic mechanisms described in the Mechanism and Catalytic Site Atlas. For template searching we optimised the geometric-matching algorithm Jess. We implemented RMSD and residue orientation filters to differentiate catalytically informative matches from spurious ones. We validated this approach on a non-redundant set of high quality experimental (n=3751, <40% amino acid identity) enzyme structures with well annotated catalytic sites as well as predicted structures of the human proteome. We show matching catalytic templates solely on structure is more sensitive than sequence- and 3D-structure-based approaches in identifying homology between distantly related enzymes. Since geometric matching does not depend on conserved sequence motifs or even common evolutionary history, we are able to identify examples of structural active site similarity in highly divergent and possibly convergent enzymes. Such examples make interesting case studies into the evolution of enzyme function. Though not intended for characterizing substrate-specific binding pockets, the speed and knowledge-driven interpretability of our method make it well suited for expanding enzyme active-site annotation across large predicted proteomes. We provide the method and template library as a Python module, Enzyme Motif Miner, at https://github.com/rayhackett/enzymm.

  bioinformatics2026-02-12v1

• Deep learning-based non-invasive profiling of tumor transcriptomes from cell-free DNA for precision oncology

  Patton, R. D.; Netzley, A.; Persse, T. W.; Nair, A.; Galipeau, P. C.; Coleman, I. M.; Itagi, P.; Chandra, P.; Adil, M.; Vashisth, M.; Sayar, E.; Hiatt, J. B.; Dumpit, R.; Kollath, L.; Demirci, R. A.; Ghodsi, A.; Lam, H.-M.; Morrissey, C.; Iravani, A.; Chen, D. L.; Hsieh, A. C.; MacPherson, D.; Haffner, M. C.; Nelson, P. S.; Ha, G.

  AI Summary

  • The study introduces Triton for fragmentomic and nucleosome profiling of cfDNA and Proteus, a deep learning framework for predicting gene expression from standard depth whole genome sequencing of cfDNA.
  • Proteus accurately reproduced gene expression profiles from ctDNA in patient-derived xenografts, similar to RNA-Seq replicates.
  • When applied to patient cohorts, Proteus predicted expression of prognostic markers, phenotype markers, and therapeutic targets, demonstrating its utility in precision oncology.

  Abstract

  Circulating tumor DNA (ctDNA) profiling from liquid biopsies is increasingly adopted as a minimally invasive solution for clinical cancer diagnostic applications. Current methods for inferring gene expression from ctDNA require specialized assays or ultra-deep, targeted sequencing, which preclude transcriptome-wide profiling at single-gene resolution. Herein we jointly introduce Triton, a tool for comprehensive fragmentomic and nucleosome profiling of cell-free DNA (cfDNA), and Proteus, a multi-modal deep learning framework for predicting single gene expression, using standard depth (~30-120x) whole genome sequencing of cfDNA. By synthesizing fragmentation and inferred nucleosome positioning patterns in the promoter and gene body from Triton, Proteus reproduced expression profiles using pure ctDNA from patient-derived xenografts (PDX) with an accuracy similar to RNA-Seq technical replicates. Applying Proteus to cfDNA from four patient cohorts with matched tumor RNA-Seq, we show that the model accurately predicted the expression of specific prognostic and phenotype markers and therapeutic targets. As an analog to RNA-Seq, we further confirmed the immediate applicability of Proteus to existing tools through accurate prediction of gene pathway enrichment scores. Our results demonstrate the potential clinical utility of Triton and Proteus as non-invasive tools for precision oncology applications such as cancer monitoring and therapeutic guidance.

  bioinformatics2026-02-12v1

• tensorOmics: Data integration for longitudinal omics data using tensor factorisation

  Kodikara, S.; Lu, B.; Wang, S.; Le Cao, K.-A.

  AI Summary

  • The study introduces tensorOmics, a framework using tensor factorization for integrating longitudinal multi-omics data, addressing the limitations of traditional matrix-based methods.
  • tensorOmics includes both supervised and unsupervised methods for single and multi-omic analyses, preserving temporal structures and integrating phenotypic responses.
  • Validation through case studies showed tensorOmics effectively differentiates treatment groups, captures time-dependent molecular signatures, and reveals coordinated responses across omics layers.

  Abstract

  Multi-omics studies capture comprehensive molecular profiles across biological layers to understand complex biological processes. A central challenge is integrating information across heterogeneous data types to identify coordinated molecular responses, particularly when measurements are collected longitudinally. Traditional integration methods can be broadly classified as unsupervised (exploring patterns without phenotypic information) or supervised (discriminating between groups or predicting outcomes). These approaches rely predominantly on matrix-based techniques that concatenate or project data into lower-dimensional spaces. However, matrix methods struggle with longitudinal data, as flattening multi-dimensional structures obscures temporal trajectories and violates independence assumptions. Tensor-based methods preserve the natural multi-way structure of longitudinal data but existing approaches are predominantly unsupervised, cannot incorporate phenotypic responses for discriminant analysis, and lack frameworks for integrating multiple omics layers. We present tensorOmics, a comprehensive framework for longitudinal omics analysis using tensor factorisation. The framework encompasses supervised and unsupervised methods for both single-omic (tensor PCA, tensor PLS discriminant analysis) and multi-omic settings (tensor PLS, block tensor PLS, block tensor PLS discriminant analysis). This unified approach captures coordinated responses across biological layers while preserving temporal structure. We validated tensorOmics through three case studies: antibiotic perturbation experiments, anaerobic digestion systems, and fecal microbiota transplantation. These applications demonstrate tensorOmics differentiates treatment groups, captures time-dependent molecular signatures, and reveals multi-layer coordinated responses that cross-sectional methods miss.

  bioinformatics2026-02-12v1

• Taxonomy-aware, disorder-matched benchmarking of phase-separating protein predictors

  Hou, S.; Shen, H.; Zhang, Y.

  AI Summary

  • The study addresses biases in existing benchmarks for phase-separating protein (PSP) predictors due to taxonomic and intrinsic-disorder imbalances.
  • A new taxonomy-aware, disorder-matched benchmark was developed, revealing that PSP features vary by taxa but LLPS-associated shifts are conserved.
  • Benchmarking 20 PSP predictors showed taxon-dependent performance variations, with PSPs lacking IDRs being particularly challenging, suggesting the need for disorder-stratified evaluations.

  Abstract

  Background: Biomolecular condensates formed via liquid-liquid phase separation (LLPS) play vital roles in cellular organization and function. Computational prediction of phase-separating proteins (PSPs) is increasingly used to prioritize candidates at proteome scale, making robust, well-designed benchmarks essential for fair evaluation and iterative improvement of PSP predictors. Results: We first show that a recently released PSP benchmark is substantially confounded by the imbalances in taxonomic origin and intrinsic-disorder compositions between positive and negative sets, allowing predictors to achieve high apparent performance by exploiting non-LLPS shortcuts and obscuring their true ability to distinguish PSPs. To minimize these effects, we construct a taxonomy-aware, disorder-matched PSP benchmark. Using this benchmark, we find that absolute sequence and biophysical feature values of PSPs differ markedly across taxa, whereas LLPS-associated feature shifts relative to taxon-specific proteome backgrounds are comparatively conserved. Benchmarking twenty PSP predictors under this framework reveals pronounced taxon-dependent variation in performance. Moreover, PSPs lacking IDRs consistently constitute a more challenging regime across methods, motivating routine disorder-stratified evaluation. Conclusions: Our taxonomy-aware, disorder-matched benchmarking framework reduces shortcut-driven biases, enables more interpretable evaluation of PSP predictors, and provides guidance for developing models that capture transferable LLPS-associated signals rather than dataset- or taxon-specific shortcuts.

  bioinformatics2026-02-12v1

• Structure-guided analysis and prediction of human E2-E3 ligase pairing specificity

  Jarboe, B.; Dunbrack, R.

  AI Summary

  • This study addresses the specificity of E2-E3 ligase interactions in ubiquitination by analyzing experimental structures from the PDB and using AlphaFold to predict thousands of ubiquitin-E2-E3 ternary complexes.
  • A machine learning model was developed to predict functional E2-E3 pairings, enhancing the understanding of ubiquitination networks.
  • The model predicted E2 partners for 88 E3 ligases, including a novel pairing between UBE2C and RNF214, potentially linking them in hepatocellular carcinoma pathways.

  Abstract

  Protein ubiquitination, directed by specific E3 ligases, constitutes the primary cellular pathway for selective protein degradation. In addition to targeting proteins for degradation, ubiquitination can mediate new protein-protein interactions, and otherwise modulate protein function, thereby regulating key cellular processes such as DNA repair and immune responses. Recently, Proteolysis-Targeting Chimeras (PROTACs), and related proximity-inducing agents, have revealed the significant therapeutic potential of co-opting ubiquitin ligase activity to induce the selective degradation of disease-relevant proteins. Despite the biological and clinical significance of this pathway, fundamental gaps remain in our understanding of ubiquitination networks, particularly regarding the specificity of E2-E3 interactions and their substrate preferences. In this study, we leverage analysis of experimental structures in the Protein Data Bank (PDB) and use AlphaFold to generate structures of thousands of ubiquitin-E2-E3 ternary complexes. Using these predicted structures and complementary analyses, we develop a machine learning model to predict functional E2-E3 pairings, advancing our ability to map ubiquitination networks and providing structural insights into functional ubiquitin-E2-E3 complexes. We demonstrate the utility of our model by predicting E2 partners for 88 putative E3 ligases lacking any previously known E2 interactors. Notably, we identify a predicted pairing between UBE2C and RNF214, two proteins recently implicated in hepatocellular carcinoma separately but through interrelated pathways, suggesting a potential functional link mediated by RNF214-dependent ubiquitination in partnership with UBE2C. Additionally, we present our web resource, UbiqCore, making the E2-E3 pairing predictions and ternary complex structures available to the scientific community (https://dunbrack.fccc.edu/ubiqcore).

  bioinformatics2026-02-12v1

• A hyperparameter benchmark of VAE-based methods for scRNA-seq batch integration

  Kassab, M.; Maniero, L.; Beltrame, E.

  AI Summary

  • The study benchmarks hyperparameters of VAE-based methods (scVI, MrVI, LDVAE) for scRNA-seq batch integration, using 960 trainings across four datasets and two feature regimes.
  • Evaluations with scib metrics showed scVI excels in batch correction, LDVAE preserves biological structure better in some datasets, and MrVI is effective in multi-protocol settings but resource-intensive.
  • Results indicated that training with highly variable genes (HVGs) generally outperformed full-gene training, and higher latent dimensionality (>30) often balanced batch mixing with biological conservation.

  Abstract

  We present the first systematic benchmark of model architecture hyperparameters for variational autoencoder (VAE) methods for single-cell RNA-seq batch integration within scvi-tools, comparing scVI, MrVI, and LDVAE across four heterogeneous datasets under two feature regimes (all genes vs highly variable genes (HVGs)). We investigated 960 trainings (120 configurations) varying latent size and network depth/width, and evaluated with a standardized scib metric suite covering batch removal and biological conservation (Batch ASW, PCR batch, iLISI, graph connectivity, NMI, ARI, label ASW, isolated-label F1/ASW, cLISI, trajectory conservation), plus qualitative UMAP/t-SNE and PCA, random projection, and unintegrated baselines. Results show dataset-dependent trade-offs: scVI performs best overall via stronger batch correction; LDVAE can better preserve biological structure in some datasets; MrVI is stable and excels at batch correction in multi-protocol settings, but is more resource-intensive. HVG-only training generally outperforms full-gene training for all models. Hyperparameter analysis suggests moderate-to-high latent dimensionality (>30) often gives the best balance; sensitivity to latent size tracks dataset heterogeneity (tissues, labs, chemistries, gene coverage), with larger latents improving batch mixing but sometimes reducing biological conservation. We provide model- and dataset-specific guidelines for practical defaults and tuning of VAE-based integration in single-cell studies.

  bioinformatics2026-02-12v1

• Splicer: Phylogenetic Placement in Sub-Linear Time

  Markin, A.; Anderson, T. K.

  AI Summary

  • Splicer is developed to perform phylogenetic placement in sub-linear time, specifically O(√n), addressing the scalability issues of existing methods like pplacer and EPA-ng.
  • It decomposes the reference tree into "blobs" and constructs a scaffold tree, then places query sequences first on the scaffold and then within blobs for precision.
  • Splicer demonstrated high accuracy on an influenza A dataset and was applied to over 12 million SARS-CoV-2 genomes, scaling maximum-likelihood placement to large datasets.

  Abstract

  Motivation: Phylogenetic placement is an established approach for rapidly classifying new genetic sequences and updating a phylogeny without fully recomputing it. Popular maximum- likelihood placement methods, such as pplacer and EPA-ng, tend to struggle computationally when the size of the reference tree increases to tens or hundreds of thousands of sequences. As a more scalable alternative, distance-based and parsimony-based placement methods were introduced such as UShER. These methods, in principle, scale linearly as the size of the reference tree grows. However, as the scale of genetic and genomic sequences continues to grow nearly exponentially, developing algorithms that can perform placement in sub-linear time while maintaining accuracy becomes more crucial. Results: Here, we develop Splicer, the first such algorithm that can perform placement in guaranteed O({surd}n) time. To achieve this performance, Splicer first decomposes the original reference tree into "blobs" and constructs a phylogenetic scaffold tree linking representatives from different blobs. Every blob in such decomposition has at most c{surd}n taxa, and the scaffold tree has at most 4/c*{surd}n leaves, where c is any constant. Then, given the query sequences for placement, they are first placed onto a scaffold tree using pplacer or EPA-ng, and then placed more precisely within the respective blobs. We demonstrate the high accuracy of Splicer on an empirical influenza A virus dataset that has sparse coverage due to limited genomic surveillance. We also show that Splicer can, for the first time, apply maximum-likelihood placement to COVID-19 pandemic-scale data using a dataset with over 12 million SARS-CoV-2 reference genomes. Splicer scales the highly accurate maximum-likelihood approaches implemented in pplacer and EPA-ng to trees with millions of taxa and eliminates the necessity to curate and subsample genomic datasets for real-time classifications. Availability and implementation: Splicer tool and source code are freely available at https://github.com/flu-crew/splicer.

  bioinformatics2026-02-12v1

• SLECA: a single-cell atlas of systemic lupus erythematosus enabling rare cell discovery using graph transformer

  Duan, M.; Shi, Y.; Tian, H.; Wu, Q.; Wang, X.; Liu, B.

  AI Summary

  • The study introduces SLECA, a large-scale single-cell atlas for systemic lupus erythematosus (SLE), using a graph-transformer framework to identify rare immune cell populations.
  • SLECA integrates 366 samples, identifying 54 cell types, including disease-relevant rare populations like double-negative T cells (DNTs), which correlate with clinical severity.
  • In silico perturbation showed that transcription factors JUN and EGR1 can reprogram DNTs, suggesting potential therapeutic targets in SLE.

  Abstract

  Systemic lupus erythematosus (SLE) is a highly heterogeneous autoimmune disease with complex immune and molecular dysregulation. While rare immune cell populations are increasingly recognized as critical drivers of disease pathogenesis and progression, the lack of sufficiently powered, comprehensive single-cell atlases has limited their systematic identification and characterization. To address this gap, we present SLECA, the first large-scale single-cell atlas of SLE, enabled by a novel graph-transformer framework for the interpretable discovery and analysis of disease-relevant rare cell populations. SLECA integrates 366 single-cell samples with standardized clinical and biological metadata, providing a comprehensive and analytically unified atlas of systemic lupus erythematosus. By enabling scalable integration and interpretable analysis, SLECA resolves 54 distinct cell types, including rare populations with critical disease relevance. Notably, we identify double-negative T cells (DNTs) as a disease-expanded population whose abundance correlates with clinical severity. Through in silico perturbation, we demonstrate that key transcription factors, specifically JUN and EGR1, can reprogram DNT cells toward conventional T-cell phenotypes, highlighting actionable regulatory vulnerabilities in SLE.

  bioinformatics2026-02-12v1

• Spatiotemporal cell type deconvolution leveraging tissue structure

  Lobo, M. M.; Zhang, Z.; Zhang, X.

  AI Summary

  • SpaDecoder is introduced as a method for cell type deconvolution in spatial transcriptomics, utilizing 3D tissue structure through an adaptive Gaussian kernel.
  • It accounts for variability in single-cell reference profiles and batch effects, enhancing the accuracy of cell type distribution estimation.
  • Comparisons and ablation tests demonstrate SpaDecoder's superior performance in leveraging 3D tissue structure for improved deconvolution across various datasets.

  Abstract

  Spot-based spatial transcriptomics (ST) captures aggregated transcriptomic profiles at spatial locations (spots) in tissue slices. Cell type deconvolution methods decode each spot and estimate the proportion of every cell type in the spot, necessary for uncovering spatial cell type distributions for further downstream analyses. Existing methods utilize cell type markers or reference transcriptomic (scRNA-seq) atlases at single cell (sc) resolution, or by aggregating profiles of identified cell types. However, current methods fail to effectively utilize the 3D tissue layout and single cell resolution reference. Some leverage 2D spatial organization assuming proximal spots are similar, which may be violated around boundaries or isolated cell types. We present SpaDecoder, a parallelized matrix factorization-based per-spot deconvolution method for multiple 3D spatial or temporal ST tissue slices effectively leveraging tissue structure with an adaptively inferred 3D neighborhood Gaussian kernel. We additionally account for variability in sc-reference profiles, along with batch effects. The mathematical framework of SpaDecoder allows it to be used for a range of downstream analyses. It can decode anteroposterior variability, impute gene expression, uncover putatively key tissue regions, identify colocalized cell types and predict spatio-temporal scRNA-seq cell locations. Ablation tests along with comparisons against other methods on various metrics, datasets, and scenarios, collectively show that SpaDecoder effectively harnesses 3D tissue structure and sc-reference profiles to improve cell type deconvolution. SpaDecoder is available at https://github.com/ZhangLabGT/spadecoder.

  bioinformatics2026-02-12v1

• MOSAIC: A Spectral Framework for Integrative Phenotypic Characterization Using Population-Level Single-Cell Multi-Omics

  Lu, C.; Kluger, Y.; Ma, R.

  AI Summary

  • MOSAIC is a spectral framework designed to analyze population-scale single-cell multi-omics data by learning a joint feature x sample embedding, addressing limitations of existing cell-centric or feature-centric methods.

  • It enables Differential Connectivity (DC) analysis, revealing regulatory network changes like the rewiring of proliferation programs in activated T cells post-vaccination, despite unchanged gene expression.

  • Applied to an HIV+ cohort, MOSAIC identified a novel stress-driven neuronal subtype with increased protein synthesis, highlighting its utility in discovering biologically significant sample subgroups.

  Abstract

  Population-scale single-cell multi-omics offers unprecedented opportunities to link molecular variation to human health and disease. However, existing methods for single-cell multi-omics analysis are either cell-centric, prioritizing batch-corrected cell embeddings that neglect feature relationships, or feature-centric, imposing global feature representations that overlook inter-sample heterogeneity. To address these limitations, we present MOSAIC, a spectral framework that learns a high-resolution feature x sample joint embedding from population-scale single-cell multi-omics data. For each individual, MOSAIC constructs a sample-specific coupling matrix capturing complete intra- and cross-modality feature interactions, then projects these into a shared latent space via spectral decomposition. The joint feature x sample embedding defines each feature's connectivity profile per sample, enabling two key downstream applications. First, MOSAIC introduces Differential Connectivity (DC) analysis, which identifies features exhibiting regulatory network rewiring across conditions even when their expression or abundance remains unchanged. Applied to a CITE-seq vaccination cohort, MOSAIC revealed rewiring of proliferation programs in activated T cells, highlighting a functional shift in STAT5B despite stable expression. Second, MOSAIC enables identification of biologically meaningful sample subgroups by isolating coherent multimodal feature modules. Applied to an HIV+ prefrontal cortex cohort, MOSAIC uncovered a novel stress-driven neuronal subtype within HIV+ samples characterized by elevated protein synthesis without chromatin accessibility changes. MOSAIC provides a general-purpose framework for systems-level phenotypic characterization, offering novel biological insights from population-scale multi-omic studies.

  bioinformatics2026-02-12v1

• Decoding the Molecular Language of Proteins with Evolla

  Zhou, X.; Han, C.; Zhang, Y.; Du, H.; Tian, J.; Su, J.; Liu, R.; Zhuang, K.; Jiang, S.; Gitter, A.; Liu, L.; Li, H.; Wu, M.; You, S.; Yuan, Z.; Ju, F.; Zhang, H.; Zheng, W.; Dai, F.; Zhou, Y.; Tao, Y.; Wu, D.; Shao, Z.; Liu, Y.; Lu, H.; Yuan, F.

  AI Summary

  • Evolla is an interactive protein-language model trained on 546 million protein-text pairs, designed to interpret protein function through natural language queries.
  • It outperforms general large language models in functional inference and matches state-of-the-art supervised models in zero-shot performance.
  • Applications include identifying eukaryotic signature proteins in Asgard archaea and discovering a novel PET hydrolase, PsPETase, validated for plastic degradation.

  Abstract

  Proteins, nature's intricate molecular machines, are the products of billions of years of evolution and play fundamental roles in sustaining life. Yet, deciphering their molecular language - understanding how sequences and structures encode biological functions - remains a cornerstone challenge. Here, we introduce Evolla, an interactive protein-language model designed to transcend static classification by interpreting protein function through natural language queries. Trained on 546 million protein-text pairs and refined via Direct Preference Optimization, Evolla couples high-dimensional molecular representations with generative semantic decoding. Benchmarking establishes Evolla's superiority over general large language models in functional inference, demonstrates zero-shot performance parity with the state-of-the-art supervised model, and exposes remote functional relationships invisible to conventional alignment. We validate Evolla through two distinct applications: identifying candidate eukaryotic signature proteins in Asgard archaea, with functional Vps4 homologs validated via yeast complementation; and interactively discovering a novel deep-sea polyethylene terephthalate (PET) hydrolase, PsPETase, confirmed to degrade plastic films. These results position Evolla not merely as a predictor, but as a generative engine capable of complex hypothesis formulation, shifting the paradigm from static annotation to interactive, actionable discovery. The Evolla online service is available at <a href="http://www.chat-protein.com/">http://www.chat-protein.com/</a>.

  bioinformatics2026-02-11v4

• scPRINT-2: Towards the next-generation of cell foundation models and benchmarks

  Kalfon, J.; Peyre, G.; Cantini, L.

  AI Summary

  • The study introduces scPRINT-2, a single-cell Foundation Model pre-trained on 350 million cells from 16 organisms, aiming to enhance performance in cell biology tasks.
  • scPRINT-2 was developed using an additive benchmark across various tasks, leading to state-of-the-art results in expression denoising, cell embedding, and cell type prediction.
  • The model's capabilities include generative functions like expression imputation and counterfactual reasoning, with demonstrated generalization to new modalities and organisms.

  Abstract

  Cell biology has been booming with foundation models trained on large single-cell RNA-seq databases, but benchmarks and capabilities remain unclear. We propose an additive benchmark across a gymnasium of tasks to discover which features improve performance. From these findings, we present scPRINT-2, a single-cell Foundation Model pre-trained across 350 million cells and 16 organisms. Our contributions in pre-training tasks, tokenization, and losses made scPRINT-2 state-of-the-art in expression denoising, cell embedding, and cell type prediction. Furthermore, with our cell-level architecture, scPRINT-2 becomes generative, as demonstrated by our expression imputation and counterfactual reasoning results. Finally, thanks to our pre-training database, we uncover generalization to unseen modalities and organisms. These studies, together with improved abilities in gene embeddings and gene network inference, place scPRINT-2 as a next-generation cell foundation model.

  bioinformatics2026-02-11v3

• Verifying LLM-extracted text with token alignment

  Booeshaghi, A. S.; Streets, A. M.

  AI Summary

  • This study investigates improving the verification of text extracted by large language models (LLMs) by aligning extracted text with the original source, focusing on discontiguous phrases.
  • Using LLM-specific tokenizers and ordered alignment algorithms, the approach improved alignment accuracy by about 50% compared to traditional word-level tokenization.
  • The study introduced the BOAT and BIO-BOAT datasets for testing, demonstrating that ordered alignment is the most practical method for this task.

  Abstract

  Large language models excel at text extraction, but they sometimes hallucinate. A simple way to avoid hallucinations is to remove any extracted text that does not appear in the original source. This is easy when the extracted text is contiguous (findable with exact string matching), but much harder when it is discontiguous. Techniques for finding discontiguous phrases depend heavily on how the text is split-i.e., how it is tokenized. In this study, we show that splitting text along subword boundaries, with LLM-specific tokenizers, and aligning extracted text with ordered alignment algorithms, improves alignment by about 50% compared to word-level tokenization. To demonstrate this, we introduce the Berkeley Ordered Alignment of Text (BOAT) dataset, a modification of the Stanford Question Answering Dataset (SQuAD) that includes non-contiguous phrases, and BIO-BOAT a biomedical variant built from 51 bioRxiv preprints. We show that text-alignment methods form a partially ordered set, and that ordered alignment is the most practical choice for verifying LLM-extracted text. We implement this approach in taln, which enumerates ordinal subword alignments.

  bioinformatics2026-02-11v2

• Augmented prediction of multi-species protein--RNA interactions using evolutionary conservation of RNA-binding proteins

  He, J.; Zhou, T.; Hu, L.-F.; Jiao, Y.; Wang, J.; Yan, S.; Jia, S.; Chen, Q.; Zhu, W.; Zhang, J.; Jia, M.; Li, Y.; Wang, X.; Wang, Y.; Yang, Y. T.; Sun, L.

  AI Summary

  • The study introduces MuSIC, a deep learning framework to predict multi-species RBP--RNA interactions by using evolutionary conservation across 11 species.
  • MuSIC outperforms existing methods, accurately predicting RBP-binding peaks with higher confidence in closely related species.
  • The framework also quantifies the impact of genetic variants on RBP binding, validated experimentally, revealing disruptions in ubiquitination pathways.

  Abstract

  RNA-binding proteins (RBPs) play critical roles in gene expression regulation. Recent studies have begun to detail the RNA recognition mechanisms of diverse RBPs. However, given the array of RBPs studied so far, it is implausible to experimentally profile RBP-binding peaks for hundreds of RBPs in multiple non-model organisms. Here, we introduce MuSIC (Multi-Species RBP--RNA Interactions using Conservation), a deep learning-based framework for predicting cross-species RBP--RNA interactions by leveraging label smoothing and evolutionary conservation of RBPs across 11 diverse species ranging from human to yeast. MuSIC outperforms state-of-the-art computational methods, and provides predicted RBP-binding peaks across species with high accuracy. The prediction confidence is higher in the closely related species, partially due to the RBP conservation patterns. Finally, the effects of homologous genetic variants on RBP binding can be computationally quantified across species, followed by experimental validations. The target transcripts with disrupted binding events are enriched with the ubiquitination-associated pathways. To summarize, MuSIC provides a useful computational framework for predicting RBP--RNA interactions cross-species and quantifying the effects of genetic variants on RBP binding, offering novel insights into the RBP-mediated regulatory mechanisms implicated in human diseases.

  bioinformatics2026-02-11v2

• Multi-compartment spatiotemporal metabolic modeling of the chicken gut guides the design of dietary interventions

  Utkina, I.; Alizadeh, M.; Sharif, S.; Parkinson, J.

  AI Summary

  • The study developed a multi-compartment, spatiotemporally resolved metabolic model of the chicken gut to understand how diet influences microbial metabolism.
  • The model identified cellulose, starch, and L-threonine as effective dietary supplements for enhancing short-chain fatty acid production, particularly butyrate, through in silico screening.
  • Validation through a feeding trial confirmed model predictions, highlighting the importance of microbial community composition in metabolic outcomes.

  Abstract

  Understanding how diet shapes microbial metabolism along the gastrointestinal tract is essential for improving poultry gut health and reducing reliance on antibiotic growth promoters. Yet dietary interventions often yield inconsistent outcomes because their efficacy depends on baseline conditions, including diet composition and microbiota structure. To address this, we developed the first multi-compartment, spatiotemporally resolved metabolic model of the chicken gastrointestinal tract. Our six-compartment framework integrates avian-specific physiological features including bidirectional flow, feeding-fasting cycles, and compartment-specific environmental parameters. The model captured distinct metabolic specialization along the gut, with upper compartments enriched for biosynthetic pathways and lower compartments specialized for fermentation. Systematic in silico screening of 34 dietary supplements revealed context-dependent metabolic responses and identified cellulose, starch, and L-threonine as robust enhancers of short-chain fatty acid production. A controlled feeding trial validated key predictions, particularly for butyrate, and integrating trial-specific microbial community data substantially improved prediction accuracy for several metabolites. Our findings demonstrate that community composition is a major driver of metabolic outcomes and underscore the need for context-specific modeling. Our framework provides a mechanistic platform for rational dietary intervention design and is broadly adaptable to other animal or human gastrointestinal systems.

  bioinformatics2026-02-11v2

• Prediction of Antibody Non-Specificity using Protein Language Models and Biophysical Parameters

  Sakhnini, L. I.; Beltrame, L.; Fulle, S.; Sormanni, P.; Henriksen, A.; Lorenzen, N.; Vendruscolo, M.; Granata, D.

  AI Summary

  • This study predicts antibody non-specificity using protein language models (PLMs) and biophysical descriptors, focusing on human and mouse antibody data.
  • The best prediction model, ESM 1v LogisticReg, achieved 71% accuracy in 10-fold cross-validation, highlighting the heavy variable domain's importance.
  • Biophysical analysis revealed the isoelectric point as a significant factor in non-specificity, with implications for developing therapeutic antibodies and nanobodies.

  Abstract

  The development of therapeutic antibodies requires optimizing target binding affinity and pharmacodynamics, while ensuring high developability potential, including minimizing non-specific binding. In this study, we address this problem by predicting antibody non-specificity by two complementary approaches: (i) antibody sequence embeddings by protein language models (PLMs), and (ii) a comprehensive set of sequence-based biophysical descriptors. These models were trained on human and mouse antibody data from Boughter et al. (2020) and tested on three public datasets: Jain et al. (2017), Shehata et al. (2019) and Harvey et al. (2022). We show that non-specificity is best predicted from the heavy variable domain and heavy-chain complementary variable regions (CDRs). The top performing PLM, a heavy variable domain-based ESM 1v LogisticReg model, resulted in 10-fold cross-validation accuracy of up to 71%. Our biophysical descriptor-based analysis identified the isoelectric point as a key driver of non-specificity. Our findings underscore the importance of biophysical properties in predicting antibody non-specificity and highlight the potential of protein language models for the development of antibody-based therapeutics. To illustrate the use of our approach in the development of lead candidates with high developability potential, we show that it can be extended to therapeutic antibodies and nanobodies.

  bioinformatics2026-02-11v2

• A multi-component power-law penalty corrects distance bias in single-cell co-accessibility and deep-learning chromatin interaction predictions

  Schlegel, L.; Gomez-Cano, F.; Marand, A. P.; Johannes, F.

  AI Summary

  • The study addresses the overestimation of long-range interactions in single-cell co-accessibility and deep learning predictions by introducing a distance-based penalty function.
  • Using Hi-C data from maize, rice, and soybean, the researchers developed tissue-specific and global consensus penalties based on multi-regime power-law exponents.
  • Applying these corrections to scATAC-seq data reduced long-range false positives by 73% with tissue-specific penalties and 66% with global consensus, aligning predictions more closely with Hi-C data.

  Abstract

  Scalable proxies for 3D genome contacts - such as single-cell co-accessibility and deep learning predictions - have emerged as powerful alternatives to chromatin capture-based methods, but predictions systematically overestimate long-range interactions. Here we show how to correct this bias using distance-based penalty functions informed by Gaussian mixture modeling and polymer-physics scaling. Using Hi-C datasets from maize, rice, and soybean, we derive tissue-specific and global consensus penalties parameterized by multi-regime power-law exponents. Applying these corrections to scATAC-seq co-accessibility scores improves their distance profiles in concordance with Hi-C and reduces long-range false positives by an average of 73% with tissue-specific penalties and 66% with the global consensus. We provide open-source code and fitted parameters to support adoption in maize, rice, and soybean.

  bioinformatics2026-02-11v2

• ModSeqR: An R package for efficient analysis of modified nucleotide data

  Zimmerman, H. E.; Moore, J.; Miller, R. H.; Stirland, I.; Jenkins, A.; Saito, E.; Jenkins, T.; Hill, J. T.

  AI Summary

  • The study addresses the computational challenges in analyzing large datasets from long-read technologies for DNA methylation.
  • They introduce the CH3 file format, reducing file sizes by over 95%, and the ModSeqR R package, which uses this format and a database backend for efficient epigenetic analyses.
  • These tools facilitate high-throughput methylation analysis with reduced computational demands.

  Abstract

  DNA methylation regulates a wide range of biological processes, including gene expression, disease progression, and cell identity. Long-read technologies now enable more comprehensive and accurate methylome analyses than ever before, but they are hindered by the computational resources needed to analyze the massive datasets. Here, we present the CH3 file format, which aids data storage and transfer by reducing file sizes by more than 95%, and the ModSeqR R package, which builds on the CH3 format and a database backend to enable a broad range of epigenetic analyses. Together, these tools enable high-throughput methylation analysis while minimizing computational resource requirements.

  bioinformatics2026-02-11v2

• Large-scale quantum computing framework enhances drug discovery in multiple stages

  Wen, K.; Zha, J.; Chen, S.; Zhong, J.; Yuan, L.; Cui, Y.; Shi, X.; Qin, W.; Lan, X.; Liu, Y.; Yang, X.; Qin, H.; Li, M.; Guo, P.; Xiao, Q.; Wu, T.; Zhou, Y.; Cao, C.; Ning, S.; Wu, C.; Gao, Q.; He, H.; Ma, Y.; An, Z.; Liu, X.; Chen, Y.; Zheng, Z.; Wei, H.; Ma, Y.; Zhang, J.

  AI Summary

  • The study improved the stability of a 2000-node Coherent Ising Machine (CIM), named QBoson-CPQC-3Gen, through enhanced vibration isolation and temperature control, allowing stable solutions for over an hour.
  • A CIM-based framework was developed for computer-aided drug discovery (CADD), incorporating graph-based encoding for tasks like allosteric site detection and protein-peptide docking.
  • This framework outperformed heuristic algorithms in speed and accuracy, identifying 2 novel druggable sites and bioactive compounds for 6 targets, validated through in vitro, in-cell, and crystallographic methods.

  Abstract

  Coherent Ising machines (CIMs) excel at solving large-scale combinational optimization problems (COPs), but their insufficient long-term stability has hindered their applications in compute-intensive tasks like computer-aided drug discovery (CADD). By improving fiber vibration isolation and temperature control system, we have implemented a 2000-node CIM named QBoson-CPQC-3Gen achieving stable solutions over one hour on large-scale COPs. Graph-based encoding schemes were further introduced to realize a CIM-based CADD workflow including allosteric site detection, protein-peptide docking and intermolecular similarity calculation. CIM-based methods demonstrated superior speed and accuracy than heuristic algorithms. Especially, QBoson-CPQC-3Gen identified 2 novel druggable sites and bioactive compounds for 6 targets, which were further validated in vitro, in-cell and by crystal structures. Our contributions established a quantum-computing framework for multi-stage drug discovery, representing a significant advancement in both quantum computing applications and pharmaceutical research.

  bioinformatics2026-02-11v1

• PlantMDCS: A code-free, modular toolkit for rapid deployment of plant multi-omics databases

  Chen, C.; Liu, Y.; Wang, L.; Sai, J.; Wang, Y.; Yue, W.; Sun, J.; Li, Z.; Wang, F.; Tian, J.; Xu, D.; Fang, Y.

  AI Summary

  • PlantMDCS is a user-friendly, code-free toolkit designed for rapid deployment of plant multi-omics databases, addressing the challenge of managing and analyzing diverse omics data.
  • It features a decoupled front-end/back-end architecture where the back end manages data storage, preprocessing, and integration, while the front end supports the entire research workflow from data import to visualization without programming.
  • Benchmarking showed that PlantMDCS can construct databases in minutes across various plant species, enhancing efficiency, reproducibility, and data security through local deployment and controlled access.

  Abstract

  With the rapid accumulation of diverse omics datasets, achieving efficient management and integrative analysis of plant multi-omics data remains a major challenge. Conventional solutions rely on constructing web-based databases, which often demand substantial programming expertise and long-term financial support. To address these limitations, we developed the Plant Multi-omics Database Construction System (PlantMDCS)-a locally deployable, user-friendly, and collaborative platform that unifies database construction and downstream multi-omics analysis within a graphical environment. PlantMDCS adopts a decoupled front-end/back-end architecture. The back end serves as the core engine for data management and computation, and is responsible for the storage, preprocessing, integration, and hierarchical association of multi-omics data. Once initialized, the front end supports the complete research workflow, including data import, querying, integrative analysis and visualization. All operations can be performed without programming, while local resource usage is dominated by disk storage required for user-provided datasets rather than sustained computational overhead. Benchmarking across plant species ranging from Arabidopsis to hexaploid wheat demonstrated that database construction can be completed within minutes, independent of genome size or data complexity. PlantMDCS is designed for local deployment to ensure data security, while allowing multi-user collaboration within local networks and supporting controlled remote access for teams distributed across different regions. Overall, PlantMDCS offers a secure and sustainable framework that integrates data management and analysis within a unified system. This design shifts multi-omics research away from fragmented file-based processing toward persistent, database-driven exploration, thereby enhancing analytical efficiency and reproducibility.

  bioinformatics2026-02-11v1

• SIPdb: A stable isotope probing database and analytical dashboard for linking amplicon sequences to microbial activity using a reverse ecology approach

  Trentin, A. B.; Simpson, A.; Kimbrel, J. A.; Blazewicz, S. J.; Wilhelm, R. C.

  AI Summary

  • SIPdb is introduced as a SQLite database and RShiny dashboard for integrating stable isotope probing (SIP) data with microbial sequence data, standardizing 22 studies across 21 isotopolog substrates.
  • The database uses a standardized pipeline to analyze SIP data, identifying over 42,000 unique amplicon sequence variants as isotope incorporators across 62 phyla, with ALDEx2 showing the highest specificity in performance.
  • Validation showed SIPdb recovered 70.1% of reported incorporator taxa, and reanalysis of a non-SIP study identified additional candidate taxa for 1,4-dioxane degradation, enhancing ecological interpretation in microbiome research.

  Abstract

  Stable isotope probing (SIP) provides a powerful means to connect microbial sequence data with diverse metabolic activities, but the lack of a framework for SIP-derived data has limited its integration into broader strategies for ecological inference. Here, we introduce the SIPdb, an extensible SQLite database of curated nucleic acid SIP experiments (also in phyloseq format) paired with an interactive RShiny dashboard for analysis and visualization. The initial release compiles 22 studies covering 21 isotopolog substrates across diverse environments, with data standardized using the MISIP metadata standard. In creating the SIPdb, we have provided a standardized pipeline that accommodates the three most common SIP gradient fractionation strategies (binary, multi-fraction, and density-resolved), two isotope incorporator designation strategies (fixed- and sliding-window), and four complementary differential abundance methods (DESeq2, edgeR, limma-voom, and ALDEx2). Using our pipeline, we identified more than 42,000 unique amplicon sequence variants as isotope incorporators across 62 phyla. Benchmarking with synthetic datasets demonstrated consistent performance across incorporator designation strategies, with ALDEx2 providing the highest specificity. Validation against original publications showed that, on average, SIPdb recovered 70.1% of author-reported incorporator taxa, with discrepancies arising from differences in phylotyping or classification approaches. Finally, our reanalysis of a non-SIP study of 1,4-dioxane degradation showed how SIPdb can both validate known degraders and uncover additional candidate taxa involved in community metabolism. The SIPdb establishes a scalable platform for reverse ecology, enabling hypothesis generation, cross-study meta-analysis, and linking taxa to metabolic processes, while serving as an open, extensible resource to accelerate ecological interpretation in microbiome research.

  bioinformatics2026-02-11v1

• DIA-CLIP: a universal representation learning framework for zero-shot DIA proteomics

  Liao, Y.; Wen, H.; E, W.; Zhang, W.

  AI Summary

  • The study introduces DIA-CLIP, a framework for zero-shot DIA proteomics that uses universal cross-modal representation learning to overcome the limitations of semi-supervised, run-specific training in DIA-MS analysis.
  • DIA-CLIP employs a dual-encoder contrastive learning approach to align peptide sequences with spectral features, enabling high-precision peptide-spectrum match inference without run-specific retraining.
  • Evaluations show DIA-CLIP increases protein identification by up to 45% and reduces false discovery rates by 17%, demonstrating superior performance over existing tools in diverse proteomic applications.

  Abstract

  Data-independent acquisition mass spectrometry (DIA-MS) has established itself as a cornerstone of proteomic profiling and large-scale systems biology, offering unparalleled depth and reproducibility.has emerged as an indispensable cornerstone of quantitative proteomics[3.1].[4.1] However, Ccurrent DIA analysis frameworks, however,identification [5.1]pipelines require semi-supervised training within each run rely on semi-supervised, run-specific training[6.1] for peptide-spectrum match (PSM) re-scoring. This approach is prone to often leads to[7.1] overfitting and lacks generalizability across diverse heterogeneous species and experimental conditionss[8.1][9.1]. Here, we present DIA-CLIP, a pre-trained modelfoundation-model-inspired[10.1] framework that shiftings the DIA analysis paradigm from semi-supervised trainingrun-specificper-file refinement to universal cross-modal representation learning.. BBy integrating dual-encoder contrastive learning framework with encoder-decoder architecture[11.1], DIA-CLIP establishes a unified cross-modal representation for peptides and corresponding spectral features, achievingemploying supervised contrastive learning on large-scale PSM datasets, DIA-CLIP aligns peptide sequences with spectral signals within a shared latent space. This approach enables high-precision, zero-shot PSM inference, eliminating the requirement for run-specific re-training or fine-tuning.[12.1] Extensive evaluations across diverse benchmarks demonstrate that DIA-CLIP consistently outperforms state-of-the-art tools, yielding up to a 45% increase in protein identification while achieving a 12% reduction in entrapment identifications.DIA-CLIP is validated to We demonstrate that DIA-CLIP consistently outperforms state-of-the-art tools across diverse benchmarks, increasing proteome coverage without compromising by 45% for single cell proteomics and reducing false discovery rates by 17% under challenging entrapment experimentfalse discovery rates under entrapment experiment[13.1]. Moreover, DIA-CLIP holds immense potential for diverse practical applications, such as single-cell and spatial proteomics, where its enhanced identification depth facilitates the discovery of novel biomarkers and the elucidates of intricate cellular mechanisms.

  bioinformatics2026-02-11v1

• VC-RDAgent: An efficient rare disease diagnosis agent via virtual case construction informed by hybrid statistical-metric and hyperbolic-semantic prioritization

  Liu, Y.; Li, H.; Jiang, P.; Wu, L.; Xie, Z.; Ning, C.; Kong, X.; Wang, Y.; Zhang, X.; Huang, Z.

  AI Summary

  • VC-RDAgent addresses the challenge of rare disease diagnosis by creating virtual standardized cases, avoiding the need for real-world patient data due to its scarcity and privacy issues.
  • The system uses VC-Ranker, which combines statistical-metric measures with hyperbolic-semantic embeddings to generate high-fidelity virtual references from knowledge bases.
  • Testing on four datasets showed VC-RDAgent improved Top-1 hit rates by 8.7% to 85.9%, with VC-Ranker achieving a Top-10 hit rate of 0.819, surpassing previous methods by 6%.

  Abstract

  While Large Language Models (LLMs) have shown promise in clinical decision support, current Retrieval-Augmented Generation (RAG) paradigms face a fundamental bottleneck in rare disease diagnosis: the scarcity, privacy restrictions, and extreme heterogeneity of real-world patient records. This reliance on sparse or inaccessible data leads to a severe "retrieval mismatch," where the lack of high-quality reference cases causes diagnostic performance to degrade sharply. To break this deadlock, we propose VC-RDAgent, a privacy-preserving and offline-capable framework that decouples diagnostic reasoning from sensitive real-world records by synthesizing virtual standardized cases. The system is powered by VC-Ranker, a multi-dimensional engine that integrates statistical-metric measures with hyperbolic-semantic embeddings to capture deep hierarchical ontology relationships. This approach allows for the dynamic generation of high-fidelity virtual references directly from authoritative knowledge bases. Extensive benchmarking across four diverse datasets demonstrates that VC-RDAgent effectively functions as a "performance equalizer." It boosts average Top-1 hit rates by 8.7% to 85.9% over zero-case baselines, enabling lightweight open-source models to rival frontier commercial systems. Notably, VC-Ranker alone achieved an aggregate Top-10 hit rate of 0.819, outperforming prior state-of-the-art methods by 6%. By eliminating the dependency on real-time web retrieval and private case sharing, VC-RDAgent provides a scalable, robust, and clinically deployable solution to shorten the diagnostic odyssey, which is made accessible through an intuitive, chat-based web application https://rarellm.service.bio-it.tech/rdagent/.

  bioinformatics2026-02-11v1

• BRIDGE: Biological Antimicrobial Resistance Inference viaDomain-Knowledge Graph Embeddings

  Iyer, A.; Kazeem, Y.; Kafaie, S.; Rajabi, E.

  AI Summary

  • The study introduces BRIDGE, a knowledge graph-based framework to enhance the prediction of antimicrobial resistance genes (ARGs) by integrating gene neighbourhood information and protein-protein interactions.
  • Focused on Klebsiella pneumoniae and Escherichia coli, BRIDGE uses data from CARD, STRING, and DrugBank to construct a knowledge graph.
  • Applying graph embedding models and deep neural networks, BRIDGE achieved a classification accuracy of up to 97% in predicting novel AMR links, demonstrating improved predictive accuracy and interpretability.

  Abstract

  Antimicrobial resistance (AMR) is a growing global health crisis, responsible for an estimated 1.27 million deaths in 2019 alone. traditional approaches to identifying antibiotic resistance genes (ARGs) are often labour-intensive and limited in their ability to detect novel resistance mechanisms. In this study, we propose BRIDGE, a knowledge graph-based framework, to improve AMR gene prediction by integrating gene neighbourhood information and protein-protein interaction networks. Focusing on Klebsiella pneumoniae and Escherichia coli, we construct a comprehensive and biologically grounded knowledge graph using curated data from CARD, STRING, and DrugBank. We apply knowledge graph embedding models which are fed into deep neural networks to infer novel AMR links, achieving classification accuracy of up to 97%. Our results demonstrate that incorporating biologically meaningful relationships, such as gene neighbourhood information and protein interactions, enhances the predictive accuracy and interpretability of AMR link predictions. This work contributes to the development of scalable and data-integrated approaches for advancing antimicrobial resistance surveillance and drug discovery.

  bioinformatics2026-02-11v1

• Siderophore identification in microorganisms associated with marine sponges by LC-HRMS and a data analytic approach in R.

  Rios, A. G.; Kato, M. J.; Yamaguchi, L. F.; Esposito, B. P.; Arenas, A. F.

  AI Summary

  • The study aimed to identify siderophores in the microbiomes of three marine sponge species using LC-HRMS and an R-based analytical workflow.
  • A total of 59 potential siderophores were annotated, with 41 confirmed through chromatographic profiling and rigorous validation criteria.
  • The approach revealed a diverse set of iron-chelating metabolites, including Ferricrocin, Aeruginic acid, and Madurastatin, without significant impact from iron supplementation during extraction.

  Abstract

  Siderophores are pivotal iron-acquisition biomolecules integral to microbial survival, pathogenicity, and ecology. Elucidating these compounds offers critical insights into the microbial dynamics of marine holobionts and potential therapeutic applications. In this study, we present a culture-independent, data-centric strategy to identify siderophores from the microbiome of three marine sponge species: Dragmacidon reticulatum, Aplysina fulva, and Amphimedon viridis. Utilizing Liquid Chromatography-High Resolution Mass Spectrometry (LC-HRMS) coupled with a custom R-based analytical workflow (XCMS and MetaboAnnotation), we successfully annotated 59 potential siderophores, 41 of which were confirmed via chromatographic profiling. We employed a rigorous validation pipeline, utilizing multiple iron-adduct calculations [M-2H+Fe]+, [M-H+Fe]2+, [2M-2H+Fe]+, high mass accuracy thresholds (<3 ppm), and retention time precision (CV < 2%). Notably, iron supplementation during extraction did not significantly alter siderophore detection, suggesting constitutive production or environmental saturation. This workflow bypasses the limitations of traditional cultivation, revealing a diverse landscape of iron-chelating metabolites--including Ferricrocin, Aeruginic acid, and Madurastatin--directly within the sponge holobiont.

  bioinformatics2026-02-11v1

• Adaptive and Spandrel-like Constraints at Functional Sites in Protein Folds

  Poley-Gil, M.; Fernandez-Martin, M.; Banka, A.; Heinzinger, M.; Rost, B.; Valencia, A.; Parra, R. G.

  AI Summary

  • The study investigates how amino acid sequences contribute to protein structure and function, focusing on the role of evolutionary and physical constraints.
  • Using reverse folding and structure prediction, researchers found that some evolutionary conserved frustration in proteins cannot be removed, suggesting these are spandrels from physical constraints.
  • These findings suggest that functional specificity in proteins might evolve from these constraints, providing insight into the interplay between evolution, structure, and biophysics.

  Abstract

  Understanding the relationships among amino acid sequences, structures and functions in proteins and how they evolve, remains a central challenge in molecular biology. It is still unclear which sequence elements differentially contribute to structural integrity or molecular function. Even more, there are ongoing debates on whether protein folds emerge as a result of evolution or as a consequence of physical laws. The energy landscapes theory states that proteins are minimally frustrated systems, i.e. they fold by minimising their energetic conflicts. However, some local frustration, believed to be selected for functional reasons, remains in the native state of proteins. Here, we combine reverse folding and structure prediction methods with sequence and local frustration analysis to address the aforementioned ideas. We found that reverse folding techniques are unable to erase evolutionary conserved frustration from certain residues, even when detrimental for structural integrity. We propose that certain frustration hotspots behave like architectural spandrels, not directly shaped by selection but emerging from physical constraints in protein folds which evolution can later co-opt for function. Our results provide a new perspective revealing how sequence variation and functional specificity could evolve from evolutionary, structural and biophysical constraints.

  bioinformatics2026-02-11v1

• BiOS: An Open-Source Framework for the Integration of Heterogeneous Biodiversity Data

  Roldan, A.; Duran, T. G.; Far, A. J.; Capa, M.; Arboleda, E.; Cancellario, T.

  AI Summary

  • The study addresses the challenge of integrating heterogeneous biodiversity data by introducing BiOS, an open-source framework designed to harmonize datasets from taxonomy, genetics, to species distribution.
  • BiOS features a modular architecture with a decoupled back-end for data management and a user-friendly front-end, offering both an API for developers and a web interface for general users.
  • Key findings include BiOS's adherence to FAIR principles, enabling seamless data integration, and enhancing collaborative conservation efforts by overcoming data fragmentation.

  Abstract

  The era of Big Data has revolutionised biodiversity research, yet the potential of this information is frequently constrained by data heterogeneity, incompatible schemas, and the fragmentation of resources. Whilst standards such as Darwin Core have improved interoperability, significant barriers persist in harmonising multi-typology datasets ranging from taxonomy and genetics to species distribution. Here, we present the Biodiversity Observatory System (BiOS), a comprehensive, open-source software stack designed to address these impediments through a modular, community-driven architecture. BiOS departs from monolithic database designs by decoupling the back-end data management from the front-end presentation layer. This architectural separation supports a dual-access model tailored to diverse stakeholder needs. For researchers and developers, the system offers a comprehensive Application Programming Interface (API) that exposes all back-end functionalities, enabling seamless programmatic access, automated data retrieval, and integration with external analytical workflows. Simultaneously, the platform features a user-centric web interface designed to lower the technical barrier to entry. This interface facilitates intuitive data exploration through agile taxonomic navigation, advanced geospatial map viewers for species occurrence filtering, and dedicated dashboards for visualising genetic markers and legislative status. Strictly adhering to the FAIR principles (Findable, Accessible, Interoperable, Reusable), BiOS acts as a relational engine capable of integrating heterogeneous data streams. By providing a flexible, interoperable core that supports the "seven shortfalls" framework of biodiversity knowledge, BiOS offers a turnkey solution to overcome data fragmentation and enhance collaborative conservation efforts.

  bioinformatics2026-02-11v1

• Cigarette smoke induces colon cancer by regulating the gut microbiota and related metabolites

  Li, W.; Bao, Y.-n.; Zhao, Q.; Yang, X.; Gong, Y.; Gan, B.

  AI Summary

  • This study investigated the link between cigarette smoke and colorectal cancer (CRC) using a mouse model, finding that smoke exposure increases CRC incidence by altering gut microbiota and related metabolites.
  • Smoke exposure decreased beneficial bacteria like Lactobacillus, increased harmful bacteria like Firmicutes and Clostridium, and altered metabolites, while also downregulating tumor suppressor genes PARG, CPT2, and ALDH1A1.
  • Functional assays confirmed that reduced CPT2 expression in CRC cells enhanced malignancy, and clinical data showed these genes were downregulated in smoking-related CRC patients, offering diagnostic potential.

  Abstract

  The causal relationship between smoking and colorectal cancer (CRC) remains unclear. In this study, a cigarette smoke-exposed mouse model demonstrated that smoking significantly increased CRC incidence by inducing gut microbiota dysbiosis and altering related metabolites. Smoke exposure reduced beneficial bacteria (e.g., Lactobacillus), increased harmful bacteria (e.g., Firmicutes and Clostridium), elevated metabolites such as histamine, and suppressed the tumor suppressor genes PARG, CPT2, and ALDH1A1, thereby promoting tumor development. Functional assays in CRC cell lines further confirmed that CPT2 knockdown enhanced malignant phenotypes, including proliferation, migration, and invasion. Clinical analysis showed that these genes were markedly downregulated in smoking-related CRC patients, with strong diagnostic value (AUC > 0.8).

  bioinformatics2026-02-11v1

• A global survey of System Biology-based predictions of gene-rare disease associations to enhance new diagnoses

  Benitez, Y.; Uria-Regojo, G.; Minguez, P.

  AI Summary

  • The study aimed to enhance rare disease diagnosis by predicting gene-disease associations using a global, network-based approach.
  • By analyzing functional neighborhoods of known disease genes, the research identified 192 genes linked to single diseases and 251 genes associated with specific disease classes.
  • These findings were used to develop a gene-disease specificity score to improve variant prioritization in genetic diagnostics.

  Abstract

  In rare disease diagnosis, described genotype-phenotype associations are evaluated first. In the absence of strong evidence, WES and WGS provide hundred to million other genetic variants, most poorly annotated, that need to be prioritized. While several in silico approaches leverage existing gene-disease knowledge to predict novel associations, doing so in isolation can hide how different genes are represented across other predictions. We hypothesize that a global perspective, accounting for differences in the knowledge accumulated in the gene collections, can refine predictions. Using a network-based algorithm, we explored functional neighborhoods of known disease-associated genes to predict novel candidates for over 200 rare diseases. A global analysis of gene and protein family behavior across predictions identified genes and functions broadly associated with multiple conditions, 192 genes linked to a single disease and 251 genes functionally associated with specific classes of rare diseases. These findings are integrated into a gene-disease specificity score, aimed at enhancing variant prioritization and guiding geneticists in advancing candidate genes toward functional validation.

  bioinformatics2026-02-11v1

• A machine learning approach to identify key Epigenetic Transcripts for Ageing research in human blood (Epitage)

  Benazzi Maia, T.; Pfeffer, U.

  AI Summary

  • This study analyzed the GSE87571 dataset to explore the relationship between transcript-level DNA methylation and chronological age in human blood, leading to the creation of Epitage.
  • Epitage consists of 48 transcripts from 13 genes, identified via machine learning, with strong age correlation (R^2 >= 0.8), including novel markers like KCNS1, SPTBN4, and VTRNA1-2.
  • An R package, ugPlot, was developed to automate model validation, enhancing reproducibility and efficiency in ageing research.

  Abstract

  DNA methylation is an established biomarker of human ageing, and analysing CpGs grouped by transcript as functional units may reveal new insights into the processes of ageing. In this study, we analyzed the GSE87571 dataset (714 samples from 14-94 years) to assess the relationship between transcript-level methylation profiles and chronological age in human blood. This approach led to the creation of Epitage, a curated set of 48 transcripts from 13 genes identified through machine learning as having methylation profiles that strongly correlate with age (R^2 >= 0.8). This analysis highlighted transcripts from the genes KCNS1, SPTBN4, and VTRNA1-2, which have been only rarely mentioned as age-related methylation markers in humans, suggesting them as underexplored candidates for future investigation. In addition, the list includes genes already implicated in aging or related pathways, such as ELOVL2, FHL2, KLF14, TRIM59, MIR29B2CHG, CALB1, OBSCN, PRRT1, OTUD7A, and SYNGR3. To validate models efficiently while ensuring reproducibility, we developed ugPlot, an open-source R package with a graphical user interface (GUI) that automates routine steps for training and testing hundreds of machine-learning models. The tool also streamlines dataset import and manipulation, reducing human error and generating publication-ready plots. Epitage thus provides a focused and accessible starting point for experimental and translational studies into the roles of DNA methylation and transcript regulation in human ageing.

  bioinformatics2026-02-11v1

• A spectral framework for measuring diversity in multiple sequence alignments

  opuu, v.

  AI Summary

  • This study introduces Leff, a spectral measure to quantify the effective diversity in multiple sequence alignments (MSAs) by estimating the number of independent positions needed to capture observed diversity.
  • Applied to RNA and protein MSAs, Leff reveals that evolutionary constraints significantly reduce diversity, with proteins showing even lower effective diversity due to stronger constraints.
  • Leff correlates with protein structure prediction accuracy and quantifies diversity in experimental libraries, serving as a tool to guide future protein and RNA design.

  Abstract

  Machine learning (ML) methods for proteins and RNAs rely on multiple sequence alignments (MSAs) and related datasets such as experimental mutagenesis libraries, yet the amount of usable information they contain remains unclear. Here, a spectral measure of information is recast into an interpretable quantity for MSAs, denoted Leff, defined as the number of fully independent alignment positions that reproduce the observed sequence diversity. Applied to RNA MSAs, this measure shows that evolutionary constraints nearly halve diversity relative to the secondary structure alone, quantifying functional and phylogenetic restrictions beyond base pairing. The same analysis indicates even lower effective diversity in proteins, quantifying stronger physicochemical and evolutionary constraints on amino acids. Leff further correlates with protein structure prediction accuracy, anticipating cases with insufficient evolutionary signal. When applied to experimentally and computationally generated libraries, it measures both produced diversity and cross-library overlap, quantifying novelty rather than redundant sampling. Together, these results establish Leff as an operational tool to estimate effective information in MSAs, anticipate modeling difficulties, and guide future protein and RNA design.

  bioinformatics2026-02-11v1

• Metadiffusion: inference-time meta-energy biasing of biomolecular diffusion models

  Lam, H. Y. I.; Pujalte Ojeda, S.; Brezinova, M.; Hanke, J.; Ong, X. E.; Mu, Y.; Vendruscolo, M.

  AI Summary

  • Metadiffusion introduces a meta-energy biasing layer to guide pretrained biomolecular diffusion models, enhancing exploration of conformational landscapes without retraining.
  • The method generates diverse conformational ensembles that align with molecular dynamics simulations, supporting optimization, targeted steering, and exploration.
  • It facilitates the study of collective variables, alternative binding poses, and ensemble generation consistent with SAXS and NMR data.

  Abstract

  Biomolecular function often depends on conformational ensembles, yet modern diffusion-based structure generators are biased toward the compact conformations prevalent in structural databases, limiting their ability to explore broad conformational landscapes. This work introduces metadiffusion, where an additional meta-energy biasing layer on top of diffusion steers pretrained biomolecular diffusion models through gradient-guided denoising. Without retraining, metadiffusion generates diverse conformational ensembles whose residue-level flexibility patterns closely match molecular dynamics simulations. The method supports three complementary modes: optimisation, steering to user-specified targets, and exploration via inter-sample repulsion. This approach enables controlled exploration of collective variables, enumeration of alternative binding poses across proteins, nucleic acids and ligands, and conformational ensemble generation consistent with SAXS and NMR chemical shifts. Metadiffusion thus provides a practical route to connect diffusion-based structure generation with ensemble-level, experimentally-restrained structural analysis.

  bioinformatics2026-02-11v1

• SCALPEL: A pipeline for processing large-scale spatial transcriptomics data

  Kunst, M.; Ching, L.; Quon, J.; Mathieu, R.; Hewitt, M.; Seeman, S.; Ayala, A.; Gelfand, E.; Long, B.; Martin, N.; Nagra, J.; Olsen, P.; Oyama, A.; Valera, N.; Pagen, C.; Sunkin, S.; Ariza, J.; Smith, K.; McMillen, D.; Zeng, H.; Waters, J.

  AI Summary

  • SCALPEL is a pipeline designed for processing large-scale spatial transcriptomics data, featuring 3D segmentation, refined filtering, doublet detection, and cell type label transfer.
  • It includes spatial domain detection and registration to the Allen Mouse Brain CCFv3, with genome-wide expression imputation from scRNAseq.
  • Benchmarking against a previous dataset showed improvements in cell number, expression clarity, and spatial registration, setting a new standard for spatial transcriptomics studies.

  Abstract

  Spatial transcriptomics enables the precise mapping of gene expression patterns within tissue architecture, offering unprecedented insights into cellular interactions, tissue heterogeneity, and disease pathology that are unattainable with traditional transcriptomic approaches. We present a tool for processing spatial transcriptomics data, SCALPEL (Spatial Cell Analysis, Labeling, Processing, and Expression Linking). SCALPEL is specifically designed to support the analysis of large, atlas-level datasets. Our new workflow features advanced 3D segmentation optimized for dense and heterogeneous tissues, refined filtering criteria, and transcriptome-based doublet detection to remove low-quality or artifactual cells. Cell type label transfer from existing taxonomies is further improved through updated filtering thresholds. Spatial domain detection is incorporated to capture local transcriptomic organization, and tissue sections are registered to the Allen Mouse Brain Common Coordinate Framework version 3 (CCFv3) for precise anatomical alignment. Genome-wide expression imputation from single-cell RNA-sequencing (scRNAseq) further enriches the dataset. Crucially, we benchmark the performance of this updated pipeline against a previously published version of our whole-mouse-brain (WMB) dataset (Yao et al., 2023b), demonstrating substantial improvements in cell number, expression profile clarity, and spatial registration. These advances provide a robust foundation for downstream spatial analyses and set a new standard for large-scale spatial transcriptomics studies.

  bioinformatics2026-02-10v3

• ETSAM: Effectively Segmenting Cell Membranes in cryo-Electron Tomograms

  Selvaraj, J.; Cheng, J.

  AI Summary

  • This study introduces ETSAM, a two-stage AI method based on SAM2, designed to segment cell membranes in cryo-ET tomograms.
  • ETSAM was trained on 83 experimental and 28 simulated tomograms, achieving state-of-the-art performance on an independent test set of 10 tomograms.
  • It significantly outperforms existing methods by providing high sensitivity and precision in membrane segmentation despite challenges like low signal-to-noise ratio and missing wedge artifacts.

  Abstract

  Cryogenic Electron Tomography (cryo-ET) is an emerging experimental technique to visualize cell structures and macromolecules in their native cellular environment. Accurate segmentation of cell structures in cryo-ET tomograms, such as cell membranes, is crucial to advance our understanding of cellular organization and function. However, several inherent limitations in cryo-ET tomograms, including the very low signal-to-noise ratio, missing wedge artifacts from limited tilt angles, and other noise artifacts, collectively hinder the reliable identification and delineation of these structures. In this study, we introduce ETSAM - a two-stage Segment Anything Model 2 (SAM2)-based fine-tuned AI method that effectively segments cell membranes in cryo-ET tomograms. It is trained on a diverse dataset comprising 83 experimental tomograms from the CryoET Data Portal (CDP) database and 28 simulated tomograms generated using PolNet. ETSAM achieves state-of-the-art performance on an independent test set comprising 10 experimental tomograms for which ground-truth annotations are available. It robustly segments cell membranes with high sensitivity and precision, significantly outperforming existing deep learning methods.

  bioinformatics2026-02-10v2

• Systems Level Analysis of Gene, Pathway and Phytochemical Associations with Psoriasis

  Ray, S.; Dutta, O.; Kousoulas, K. G.; Apostolopoulos, N.; Chamcheu, J. C.; Kaur, R.

  AI Summary

  • The study used a systems biology approach to analyze gene expression and pathways in psoriatic lesions, identifying key roles of type I/III interferon signaling, AP-1, and CREB1.
  • It highlighted seven phytochemicals with potential multi-target activity against psoriasis, focusing on the IL-17/TNF-interferon-AP-1/CREB1-COX-2/MMP9 axis.
  • Protopine and atractylon were suggested as promising candidates for topical treatment due to favorable ADMET properties, with further validation needed in skin models.

  Abstract

  Psoriasis is an inflammatory skin disorder driven by abnormal immune activation that promotes excessive proliferation and accelerated turnover of epidermal keratinocytes. IL-17 and TNF pathways are well known in psoriasis, but the other mechanisms that keep the disease active and link it to systemic comorbidities are not yet fully understood. A combined transcriptomic and systems biology framework was applied to map regulatory circuits in psoriatic lesions and to identify phytochemical candidates capable of multi-target modulation for topical intervention. Differential gene expression between lesional and healthy skin was analyzed, followed by pathway enrichment, upstream regulator inference, protein-protein interaction network, and chemical-gene interaction mapping. This integrative strategy revealed a transcriptional landscape dominated by type I/III interferon signaling, antiviral and antimicrobial responses, immune metabolic dysregulation, and transcriptional hubs centered on AP-1 and CREB1. Several genes and upstream regulators not previously associated with psoriasis were identified within inflammatory and cell migration-related modules, indicating unexplored regulatory layers in disease control. Network-guided chemical prioritization and direction-of-effect filtering highlighted seven phytochemicals (mahanine, atractylon, protopine, annomontine, taraxasterol, tricin, and tamarixetin) with multi-target activity across key disease axes. ADMET-based screening suggested protopine and atractylon as favorable candidates for topical delivery, while synergy modeling supported flavonoid-alkaloid combination designs. This multi-layered approach provides mechanistically informed phytochemicals targeting the IL-17/TNF-interferon-AP-1/CREB1-COX-2/MMP9 axis in psoriasis. Experimental validation in keratinocyte and organotypic skin models will be required to determine whether these compounds, individually or in combination, can effectively restore psoriatic signaling in vivo.

  bioinformatics2026-02-10v2

• Autoregressive forecasting of future single-cell state transitions

  Luo, E.; Gao, H.; BIAN, H.; Li, Y.; Li, C.; Hao, M.; Chen, M.; She, Y.; Wei, L.; Liu, K.; Zhang, X.

  AI Summary

  • The study introduces CellTempo, a temporal generative AI model designed to forecast future cellular dynamics from static single-cell RNA-sequencing data.
  • CellTempo uses learned semantic codes and an autoregressive decoder to predict long-range cell-state transitions.
  • Experiments demonstrated that CellTempo accurately forecasts cell state evolutions and reconstructs cell-state landscapes post-perturbations, aligning well with biological realities.

  Abstract

  Existing methods for dynamic analysis of static single-cell RNA-sequencing data can reconstruct temporal structures covered by observed cells, but cannot forecast unobserved future state transitions. We propose a temporal generative AI model, CellTempo, to forecast future cellular dynamics by representing cells as learned semantic codes and training an autoregressive generation decoder to predict ordered code sequences. It can forecast long-range cell-state transition trajectories and landscapes from snapshot data. To train the model, we constructed a comprehensive single-cell trajectory dataset scBaseTraj by integrating RNA velocity, pseudotime, and inferred transition probabilities to compose multi-step cellular sequences. Experiments on multiple real datasets showed that CellTempo can forecast cell state evolutions from individual cells, and reconstruct nuanced cell-state potential landscapes and their varied progressions after genetic or chemical perturbations, all with high fidelity to biological truth. This work opens a route for forecasting unseen future dynamics of cell state transitions from static observations.

  bioinformatics2026-02-10v1

• Optimizing Protein Tokenization: Reduced Amino Acid Alphabets for Efficient and Accurate Protein Language Models

  Rannon, E.; Burstein, D.

  AI Summary

  • This study explores the use of reduced amino acid alphabets combined with Byte Pair Encoding (BPE) tokenization in protein language models (pLMs) to optimize efficiency.
  • RoBERTa-based pLMs were pre-trained using various reduced alphabets and evaluated on multiple tasks.
  • Results indicated that reduced alphabets significantly shortened input sequences, sped up training and inference, and maintained or improved performance compared to models using the full 20-amino-acid alphabet.

  Abstract

  Protein language models (pLMs) typically tokenize sequences at the single-amino-acid level using a 20-residue alphabet, resulting in long input sequences and high computational cost. Sub-word tokenization methods such as Byte Pair Encoding (BPE) can reduce sequence length but are limited by the sparsity of long patterns in proteins encoded by the standard amino acid alphabet. Reduced amino acid alphabets, which group residues by physicochemical properties, offer a potential solution but their performances with sub-word tokenization have not been systematically studied. In this work, we investigate the combined use of reduced amino acid alphabets and BPE tokenization in protein language models. We pre-trained RoBERTa-based pLMs de novo using multiple reduced alphabets and evaluated them across diverse downstream tasks. Our results show that reduced alphabets enable substantially shorter input sequences and faster training and inference, while maintaining comparable, and in some cases improved, performance relative to models trained on the full 20-amino-acid alphabet. These findings demonstrate that alphabet reduction facilitates more effective sub-word tokenization and provides a favorable trade-off between efficiency and predictive accuracy.

  bioinformatics2026-02-10v1

• PEhub resolves the hierarchical regulatory architecture of multi-way enhancer hubs in the human brain

  Tan, J.; Sun, Y.

  AI Summary

  • PEhub is a new framework that resolves multi-way enhancer hubs from chromatin interaction data by modeling synergistic enhancer cooperation and accounting for interaction decay.
  • Using H3K27ac HiChIP data, PEhub identified and validated promoter-anchored enhancer hubs in six human brain regions, showing they correspond to real multi-way chromatin assemblies.
  • Enhancer hubs were found to be associated with increased transcription, hierarchical organization, and linked to genetic risk and transcription factor deployment in brain regions.

  Abstract

  Chromatin interaction assays capture regulatory architecture as stochastic pairwise contacts, limiting the ability to resolve how multiple enhancers cooperatively regulate transcription. Here we introduce a promoter-centric quantitative framework, termed PEhub, that resolves multi-way enhancer hubs as higher-order regulatory units from chromatin interaction data. By reparameterizing stochastic pairwise ligation events into promoter-conditioned enhancer networks, our approach explicitly models synergistic enhancer cooperation while accounting for distance-dependent interaction decay through a statistically principled null model. Using H3K27ac HiChIP data, we identify promoter-anchored enhancer hubs and validate their physical existence with single-molecule Pore-C, demonstrating that inferred hubs correspond to bona fide multi-way chromatin assemblies. Application to six human brain regions reveals that enhancer hubs are associated with elevated transcriptional output and exhibit a hierarchical organization spanning shared, circuit-specific, and region-restricted regulatory programs. This architecture hierarchically stratifies genetic risk and transcription factor deployment, linking three-dimensional genome organization to transcriptional control and disease-associated variation. Together, this promoter-centric framework provides a generalizable strategy for resolving higher-order regulatory architecture from 3D genome data and establishes multi-way enhancer hubs as a functionally and genetically meaningful layer of transcriptional regulation in complex tissues.

  bioinformatics2026-02-10v1

• Token Alignment for Verifying LLM-Extracted Text

  Booeshaghi, A. S.; Streets, A. M.

  AI Summary

  • The study investigates improving the verification of text extracted by large language models (LLMs) by aligning extracted text with the original source, focusing on discontiguous phrases.
  • Using LLM-specific tokenizers and ordered alignment algorithms, the approach improved alignment accuracy by about 50% over word-level tokenization.
  • The effectiveness was demonstrated with the introduction of the BOAT and BIO-BOAT datasets, showing ordered alignment as the most practical method for this task.

  Abstract

  Large language models excel at text extraction, but they sometimes hallucinate. A simple way to avoid hallucinations is to remove any extracted text that does not appear in the original source. This is easy when the extracted text is contiguous (findable with exact string matching), but much harder when it is discontiguous. Techniques for finding discontiguous phrases depend heavily on how the text is split-i.e., how it is tokenized. In this study, we show that splitting text along subword boundaries, with LLM-specific tokenizers, and aligning extracted text with ordered alignment algorithms, improves alignment by about 50% compared to word-level tokenization. To demonstrate this, we introduce the Berkeley Ordered Alignment of Text (BOAT) dataset, a modification of the Stanford Question Answering Dataset (SQuAD) that includes non-contiguous phrases, and BIO-BOAT a biomedical variant built from 51 bioRxiv preprints. We show that text-alignment methods form a partially ordered set, and that ordered alignment is the most practical choice for verifying LLM-extracted text. We implement this approach in taln, which enumerates ordinal subword alignments.

  bioinformatics2026-02-10v1

• bMINTY: Enabling Reproducible Management of High-Throughput Sequencing Analysis Results and their Metadata

  Kapelios, K.; Xiropotamos, P.; Manousaki, H.; Sinnis, C.; Kotsira, V.; Dalamagas, T.; GEORGAKILAS, G. K.

  AI Summary

  • The study addresses the challenge of managing high-throughput sequencing data by introducing bMINTY, a web application for structured management of post-alignment data and metadata.
  • bMINTY allows for the integration of study, assay, and analysis metadata into a single, portable, queryable resource, enhancing data reuse and reproducibility.
  • Users can export data in RO-Crate format, facilitating machine-readable data packages for publication, thereby promoting FAIR science principles.

  Abstract

  Due to the large scale of high-throughput sequencing data generation, the community and publishers have established standards for the dissemination of studies that produce and analyze these data. Despite efforts towards Findable, Accessible, Interoperable and Reproducible (FAIR) science, critical obstacles remain. Best practices are not consistently enforced by scientific publishers, and when they are, essential information is fragmented across the methods section, supplementary materials, and public repositories. When attempting to reproduce scientific findings or reuse published data or analyses, researchers often avoid analyzing sequencing data from the ground up. Instead, they prefer to start directly from the post-sequence-alignment information (e.g., gene expression matrices in transcriptomics). However, existing repositories and workflow-oriented solutions rarely provide a single, portable, queryable resource that integrates this information with the metadata required for downstream reuse. We introduce bMINTY, a locally deployed web application with an intuitive user interface, for structured management of post-alignment workflow data outputs. bMINTY supports metadata for studies, assays, and analysis assets, including workflows, genome assemblies, genomic intervals, and cell-level entities for single-cell assays. Users may export query results in RO-Crate format, providing machine readable data packages and metadata. To the best of current knowledge, bMINTY is the first framework to bundle all this information in publication-ready, portable packaging designed for reuse. These packages can be included as supplementary material with each publication, accompanied by analysis code deposited in public repositories for downstream ad hoc analyses. Together, these practices can promote transparency, efficient reuse of published data, and support FAIR-aligned scientific reproducibility.

  bioinformatics2026-02-10v1

• SenNet Portal: Build, Optimization and Usage

  Borner, K.; Blood, P. D.; Silverstein, J. C.; Ruffalo, M.; Satija, R.; Gehlenborg, N.; Honick, B.; Bueckle, A.; Jain, Y.; Qaurooni, D.; Shirey, B.; Sibilla, M.; Metis, K.; Bisciotti, J.; Morgan, R. S.; Betancur, D.; Sablosky, G. R.; Turner, M. L.; Kim, S.-J.; Lee, P. J.; Bartz, J.; Domanskyi, S.; Peters, S. T.; Enninful, A.; Farzad, N.; Fan, R.; SenNet Team, ; Herr, B. W.

  AI Summary

  • The SenNet Program addresses the challenge of studying cellular senescence by generating multimodal datasets across human and mouse tissues.
  • The SenNet Data Portal provides open access to these datasets, including single-cell, spatial, imaging, transcriptomic, and proteomic data, along with senescence biomarker catalogs and standardized protocols.
  • The portal, built on a scalable hybrid cloud architecture, supports data submission, analysis, and cross-species mapping, with applications in biomarker discovery and spatial analysis.

  Abstract

  Cellular senescence is a hallmark of aging and a driver of functional decline across tissues, yet its heterogeneity and context dependence have limited systematic study. The Common Fund Cellular Senescence Network (SenNet) Program addresses this challenge by generating multimodal, multi-tissue datasets that profile senescent cells across the human lifespan and complementary mouse models. The SenNet Data Portal (https://data.sennetconsortium.org) serves as the public gateway to these resources, providing open access to harmonized single-cell, spatial, imaging, transcriptomic, and proteomic data; senescence biomarker catalogs; and standardized protocols that can be used to comprehensively identify and characterize senescent cells in mouse and human tissue. As of January 2026, the portal hosts 1,753 publicly available human and mouse datasets across 15 organs using 6 general assay types. Experts from 13 Tissue Mapping Centers (TMCs) and 12 Technology Development and Application (TDAs) components contribute tissue data, analyze data, identify senescent biomarkers, and agree on panels for cross-tissue antibody harmonization. They also register human tissue data into the Human Reference Atlas (HRA) and develop user interfaces for the multiscale and multimodal exploration of this data. Built on a scalable hybrid cloud microservices architecture by the Consortium Organization and Data Coordinating Center (CODCC), the Portal enables data submission, management, integrated analysis, spatial context mapping, and cross-species senescence mapping critical for aging research. This paper presents user needs, the Portal architecture, data processing workflows, and senescence-focused analytical tools. The paper also presents usage scenarios illustrating applications in biomarker discovery, quality benchmarking, hypothesis generation, spatial analysis, cost-efficient profiling, and cell distance distribution analysis. Current limitations and planned extensions, including expanded spatial-omics releases and improved tools for senotype characterization, are discussed. SenNet protocols, code, and user interfaces are freely available on https://docs.sennetconsortium.org/apis.

  bioinformatics2026-02-10v1

• Using user-centered design to better understand challengesfaced during genetic analyses by novice genomicresearchers

  Patel, H.; Crosslin, D.; Jarvik, G. P.; Hall, T.; Veenstra, D.; Xie, S.

  AI Summary

  • This study aimed to understand the challenges novice genomic researchers (NGRs) face with bioinformatics tools by using a user-centered design approach.
  • A literature review and semi-structured interviews were conducted to identify issues like poor documentation, installation difficulties, and unclear error messages.
  • An evaluation rubric was developed to assess bioinformatics tools, aiming to improve usability for both NGRs and experienced users.

  Abstract

  The lack of user-centered design principles in the current landscape of commonly-used bioinformatics software tools poses challenges for novice genomics researchers (NGRs) entering the genomics ecosystem. Comparing the usability of one analysis software to that of another is a non-trivial task and requires evaluation criteria that incorporates perspectives from both existing literature and a diverse, underrepresented user base of NGRs. To better characterize these barriers, we utilized a two-pronged approach consisting of a literature review of existing bioinformatics tools and semi-structured interviews of the needs of NGRs. From both knowledge sources, the key attributes that resulted in poor adoption and sustained use of most bioinformatics tools included poor documentation, lack of readily-accessible informational content, challenges with installation and dependency coordination, and inconsistent error messages/progress indicators. Combining the findings from the literature review and the insights gained by interviewing the NGRs, an evaluation rubric was created that can be utilized to grade existing and future bioinformatics tools. This rubric acts as a summary of key components needed for software tools to cater to the diverse needs of both NGRs and experienced users. Due to the rapidly evolving nature of genomics research, it becomes increasingly important to critically evaluate existing tools and develop new ones that will help build a strong foundation for future exploration.

  bioinformatics2026-02-10v1

• PRIZM: Combining Low-N Data and Zero-shot Models to Design Enhanced Protein Variants

  Harding-Larsen, D.; Lax, B. M.; Garcia, M. E.; Mendonca, C.; Mejia-Otalvaro, F.; Welner, D. H.; Mazurenko, S.

  AI Summary

  • PRIZM is a two-phase workflow that uses a small, high-quality dataset to select the best pre-trained zero-shot model for predicting protein variant effects.
  • It then applies this model to rank and prioritize variants for experimental testing.
  • In case studies, PRIZM improved enzyme variants, achieving a 3°C increase in thermostability and a 20% increase in activity.

  Abstract

  Machine learning has repeatedly shown the ability to accelerate protein engineering, but many approaches demand large amounts of robust, high-quality training data as well as substantial computational expertise. While large pre-trained models can function as zero-shot proxies for predicting variant effects, selecting the best model for a given protein property is often non-trivial. Here, we introduce Protein Ranking using Informed Zero-shot Modelling (PRIZM), a two-phase workflow that first uses a high-quality low-N dataset to identify the most suitable pre-trained zero-shot model for a target protein property and then applies that model to rank and prioritize an in silico variant library for experimental testing. Across diverse benchmark datasets spanning multiple protein properties, PRIZM reliably separated low- from high-performing models using datasets of ~20 labelled variants. We further demonstrate PRIZM in enzyme engineering case studies targeting sucrose synthase thermostability and glycosyltransferase activity, where PRIZM-guided selection identified improved variants, including gains of ~3{degrees}C in apparent melting temperature and ~20% higher relative activity. PRIZM provides an accessible, data-efficient route to leverage foundation models for protein design while requiring minimal experimental data.

  bioinformatics2026-02-10v1

• An Integrated Pipeline for Cell-Type Annotation, Metabolic Profiling, and Spatial Communication Analysis in the Liver using Spatial Transcriptomics

  Zhang, C.; Li, J.; Luo, O.; Andrews, T.; Steinberg, G. R.; WANG, D.

  AI Summary

  • The study presents a protocol for analyzing spatial transcriptomics (ST) data in liver tissues from MASLD mouse models to understand liver metabolism.
  • The approach includes single-cell RNA-seq referencing, manual annotation with curated liver cell type markers, and metabolic gene set analysis.
  • Key findings include the provision of tools for researchers to decode metabolic reprogramming and cellular heterogeneity in liver health and disease.

  Abstract

  The liver acts as a central metabolic hub, integrating systemic signals through a spatially organized pattern known as zonation, driven by the coordinated activity of diverse cell types including hepatocytes, stellate cells, Kupffer cells, endothelial cells, and immune populations. Spatial transcriptomics (ST) enables the profiling of thousands of cells with spatial resolution in a single experiment, facilitating the identification of novel gene markers, cell types, cellular states, and tissue neighborhoods across diverse tissues and organisms. By simultaneously capturing transcriptional and spatial heterogeneity, ST has become a powerful tool for understanding cellular and tissue biology. Given its advantages, there is growing demand for applying ST to uncover novel biological insights in the liver under various physiological and pathological conditions including obesity, diabetes, and metabolic dysfunction-associated steatotic liver disease (MASLD). However, to date no comprehensive and practical protocols currently exist for analyzing ST data specifically in the context of liver metabolism. Herein, we present a systematic and detailed protocol for ST data analysis using liver tissues from MASLD mouse models. This guide offers practical support for metabolic based researchers without advanced expertise in coding, mathematics and statistics enabling single-cell RNA-seq referencing for deconvolution-based annotation, curated liver cell type markers for manual annotation, and a GMT file of metabolic gene sets and flux balance analysis to analyze liver metabolic activity. This framework and integrated computational resources for decoding metabolic reprogramming and cellular heterogeneity will empower researchers to uncover novel biological pathways regulating liver metabolism in health and disease.

  bioinformatics2026-02-10v1

• HORDCOIN: A Software Library for Higher Order Connected Information and Entropic Constraints Approximation

  Raffaelli, G. T.; Kislinger, J.; Kroupa, T.; Hlinka, J.

  AI Summary

  • The study introduces HORDCOIN, a software library for approximating higher-order connected information in complex systems like neuronal populations, using an entropic-constraint approach to simplify computational complexity.
  • This method transforms the problem into a linear program, allowing efficient estimation even with limited data.
  • Applications to symbolic sequences, neuronal recordings, and DNA sequences showed accurate detection of higher-order interactions, demonstrating the library's utility in biomedical data analysis.

  Abstract

  Background and objective: Quantifying higher-order statistical dependencies in multivariate biomedical data is essential for understanding collective dynamics in complex systems such as neuronal populations. The connected information framework provides a principled decomposition of the total information content into contributions from interactions of increasing order. However, its application has been limited by the computational complexity of conventional maximum entropy formulations. In this work, we present a generalised formulation of connected information based on maximum entropy problems constrained by entropic quantities. Methods: The entropic-constraint approach, contrasting with the original constraints based on marginals or moments, transforms the original nonconvex optimisation into a tractable linear program defined over polymatroid cones. This simplification enables efficient, robust estimation even under undersampling conditions. Results: We present theoretical foundations, algorithmic implementation, and validation through numerical experiments and real-world data. Applications to symbolic sequences, large-scale neuronal recordings, and DNA sequences demonstrate that the proposed method accurately detects higher-order interactions and remains stable even with limited data. Conclusions: The accompanying open-source software library, HORDCOIN (Higher ORDer COnnected INformation), provides user-friendly tools for computing connected information using both marginal- and entropy-based formulations. Overall, this work bridges the gap between abstract information-theoretic measures and practical biomedical data analysis, enabling scalable investigation of higher-order dependencies in neurophysiological and other complex biological systems such as the genome.

  bioinformatics2026-02-10v1