Bioinformatics

MotivationConstraint-based metabolic modeling faces a calibration gap: genome-scale metabolic models (GEMs) integrated with transcriptomics alone rely on expression-to-flux heuristics (E-Flux, GIMME, iMAT, MOMENT) that ignore cross-omics co-variance structure and lack statistical mechanisms for propagating omics uncertainty into reaction bounds, yielding flux predictions with limited agreement to 13C metabolic flux analysis (MFA) measurements. ResultsWe present Chemo-Calib, a multiblock PLS (MB-PLS) framework that calibrates GEM reaction bounds from the shared latent structure of metabolomics, transcriptomics, and proteomics data. On 11 E. coli 13C-MFA reference conditions spanning the Keio fluxome and Holm 2010 datasets, ChemoCalib constrained FBA on iJO1366 achieves a Spearman{rho} = 0.461 overall (up to 0.523 in PPP) and Pearson r of 0.49-0.58 across central carbon pathways, with statistically significant improvement over expression-only baselines including E-Flux2 and SPOT (p < 0.05, Holm-corrected). The latent-to-constraint mapping employs GPR-aware VIP aggregation (Algorithm 1) to project multi-omics latent scores onto genome-scale reaction bounds without heuristic thresholding. An optional in-silico active learning loop (relegated to Supplementary Material) further tightens calibration through virtual experiment selection. AvailabilityChemoCalib is open-source (MIT) at https://github.com/chemocalib/chemocalib with Docker support, a 5-minute tutorial, and pre-computed iJO1366 benchmarks. Preprint available at bioRxiv; code archived at Zenodo DOI: 10.5281/zenodo.21645890.

MotivationHigh-resolution spatial transcriptomics enables single-cell ligand-receptor analysis, but unmasked receptor spatial lags include receptor expression from non-target neighbors, complicating the attribution of local communication signals to specified source-target cell-type pairs. ResultsWe present MaskTalk, a Python package implementing the cell-identity-gated spatial lag model (CIG-SLM). CIG-SLM restricts receptor-side neighborhoods to target cells through W(C) = WD(C). In cell-level breast cancer Visium HD data, CIG-SLM produced target-cell-dependent communication profiles relative to the matched LIANA+ bivariate unmasked baseline, and masked-specific records showed larger between-condition effect sizes and shorter physical source-target distances. Public breast cancer Xenium data demonstrated that MaskTalk runs on external cell-level spatial data and exhibits target-aware masking behavior. Availability and ImplementationImplemented in Python with AnnData; code is available at https://github.com/JiaPP1994/MaskTalk. Software and data archive DOIs are 10.5281/zenodo.21735883 and 10.5281/zenodo.21735951, respectively. Contactshaobin79@aliyun.com; cuihw2001423@163.com Supplementary InformationSupplementary Methods S1-S4, Figures S1-S3, and Tables S1-S5.

Summary: Automated protein function annotation remains challenging as sequence databases outpace curated labels and homology-based transfer fails for proteins lacking close relatives. We present GOlien, a scalable, composition-based method that annotates protein sequences using a Shannon-entropy k-mer model. Built from the CAFA3-based training split and evaluated on the held-out validation split, GOlien achieves micro-averaged of 0.7317 (Biological Process), 0.7610 (Cellular Component) and 0.8295 (Molecular Function). Our tool offers a practical, complementary alternative to existing pipelines and extending annotation coverage for proteins. Availability and Implementation: The command line tool to submit FASTA files and retrieve predicted GO term annotations with the corresponding documentation is available at: https://github.com/probalytiq/golien-tool

SummaryA modular FASTQ-to-figures solution for analyzing low-depth or shallow whole-genome sequencing (sWGS) data. Shallow WGS can be used to detect copy number (CN) aberrations, Homologous Recombination Deficiency (HRD), and to detect and create CN signatures. Growing in popularity, this sequencing type is used for neonatal diagnostics and studying cancer. One of the major benefits is the reduced cost compared with deeper sequencing modalities such as Whole Genome Sequencing (WGS). With just 15 million reads often targeted, and sample substrate options like Formalin-Fixed, Paraffin-Embedded (FFPE) blocks widely available, this approach enables an affordable study to be performed at scale. Our pipeline and R package are an end-to-end solution implemented with reusability and modularity in mind. It makes entry and exit from the ecosystem easy, providing regular standardized output formats throughout execution. The pipeline is written in the well-supported and cross-platform Nextflow framework and has been submitted for inclusion in nf-core. Additionally, a Docker image for the utanos R package has been created to improve modularity. Availability and ImplementationThe latest version of all software is freely available on GitHub. For the full processing pipeline, visit: https://github.com/Huntsmanlab/swgs-processing-pipeline. For just the utanos R package, visit: https://github.com/Huntsmanlab/utanos.

SummaryElectronic health record (EHR)-linked biobanks generate unprecedented genomic and phenotypic datasets, but their scientific utility is constrained by data fragmentation across institutional silos and incompatible computing infrastructures, forcing researchers to rewrite ad-hoc scripts for each new environment. We present the PMBB Geno-Pheno Toolkit, a suite of modular Nextflow pipelines for biobank-scale association analyses. This note focuses on the toolkits SAIGE family of pipelines -- supporting genome-wide (GWAS), exome-wide (ExWAS), and phenome-wide (PheWAS) association testing -- together with the companion GWAMA and ExWAS meta-analysis pipelines that enable cross-biobank replication. All components are containerized (Docker/Apptainer) and orchestrated with Nextflow, allowing the same workflows to run unmodified on local HPC clusters, cloud platforms, and the All of Us Research Workbench. Complementary toolkit pipelines for PLINK-based GWAS, polygenic scoring, LD-based clumping, and phenotype harmonization are also available and briefly noted. AvailabilityThe PMBB Geno-Pheno Toolkit is freely available at https://github.com/PMBB-Informatics-and-Genomics/pmbb-geno-pheno-toolkit under MIT open-source license.

MotivationKinases regulate a multitude of protein functions, and their dysregulation is pivotal for many human diseases. Direct measurement of kinase activity, however, is often challenging; therefore, inferring activity from the behavior of their substrates is a widely adopted strategy. Nonetheless, traditional methods typically oversimplify the underlying network, ignoring that any particular substrate can be phosphorylated by multiple kinases. ResultsWe present LIKA, a likelihood-based framework for inferring kinase activity from phosphoproteomic data. By modeling the many-to-many structure of kinase-substrate interactions, LIKA achieves high efficiency, even with limited data, while capturing network complexity. Simulation and cell line analyses confirm the robustness and accuracy of LIKA. Importantly, analysis of a phosphoproteomic dataset from schizophrenia and control subjects reveals novel dysregulated kinases. Availability and ImplementationThe implementation code and publicly available data are provided at: https://github.com/lujingz/LIKA.

MotivationLigand binding-site similarity search is a crucial step in drug discovery that reduces the conformational search space for docking and other downstream tasks by comparing a target protein against experimentally identified binding sites. Existing methods rely on either direct structural alignment or lossy compression of structural information, producing a trade-off between scalability and precision. ResultsWe propose LEN-Seek, a ligand binding-site search method based on a graph neural network (GNN)-driven variational autoencoder (VAE) that encodes the 3D structural and physicochemical context of a binding site into a probabilistic latent space, enabling similarity search within a low-dimensional vector space. A binding site is modeled as a graph of amino acid residues, with node features adopted from the protein language model, Ankh, and edges encoded as SE(3)-invariant (roto-translational invariant) geometric relationships, thereby avoiding expensive data augmentation or SE(3)-equivariant models. Compared to ProBiS, the purely geometric graph-clique based method, LEN-Seek successfully retrieves a substantial portion of similar binding sites with a roughly 3,400-fold lower per-comparison cost, demonstrating its potential as a scalable approach to template-based ligand binding-site search in large-scale protein structure databases. Supplementary informationSupplementary data are available at Bioinformatics online.

We describe LRSPAT (low-rank spatial toolkit), a fast and memory-efficient framework for approximating measures of spatial association for high-dimensional data. While LRSPAT can be applied to any multivariate spatial dataset, development was motivated by the computational challenge of identifying spatially variable genes in high-resolution spatial transcriptomics (ST) data generated by technologies such as 10x Visium HD, Xenium and Atera. LRSPAT leverages a truncated SVD of the expression data and a thresholded spatial weights matrix to perform reduced-rank reconstruction of spatial statistics in the quadratic form family, including global and local versions of Moran's I, Geary's C, and Getis-Ord G. A regularization approach is leveraged to account for the inflated null distribution of spatial statistics computed on latent variables. By performing key operations on the low-dimensional embeddings, LRSPAT is orders of magnitude faster than standard implementations with significantly lower memory requirements. Because the low-rank approach denoises and desparsifies ST data, LRSPAT is also more accurate than standard techniques at identifying genes with true spatial expression patterns. The dramatic improvements in execution time and memory consumption enable the genome-wide analysis of spatially variable genes (SVGs) and exploration of the full range of hyperparameters including spatial scale, distance metric, and embedding rank. This preprint outlines the background and mathematical details of the approach with limited preliminary results and a short conclusion.

MotivationDifferential gene expression analysis (DEA) via RNA sequencing (RNA-seq) is essential but remains challenging for wet-lab biologists due to command-line complexities. Centralized web platforms democratize this process but suffer from server congestion, long queuing delays, data privacy risks with proprietary datasets, and limited long-term sustainability due to hosting fees. ResultsWe present DEAR-OWL (Differential Expression Analysis Resource on the Web (Lite)), a fully serverless, privacy-preserving web application that performs the DEA locally inside the users web browser. To combine instant exploratory speed with rigorous verification, the application runs two distinct analysis options. The first option is a fast screening tool written in native browser language (JavaScript) that delivers immediate, genome-wide fold-change calculations and statistical screening based on an edgeR-equivalent logic. The second option is a heavy-duty statistical tool that brings the standard R package (DESeq2) directly into the browser using WebR and WebAssembly technology, ensuring publication-grade validation without needing server power. Interactive visual plots (volcano plots, minus-average plots, and heatmaps) are seamlessly generated from the results of either analysis choice. Benchmarking proved its hardware compatibility: the browser-based DESeq2 engine completed the analysis in [~]30 seconds on a 64 GB RAM workstation and in [~]3 minutes on an 8 GB RAM laptop without crashing. DEAR-OWL can utilize the Grass Expression Atlas (GExA) data as built-in and supports secure local file uploads, ensuring total data privacy with neither queuing delays nor cloud infrastructure costs. Availability and implementationDEAR-OWL is freely accessible at https://webpark2116.sakura.ne.jp/deseq2/. The source code is available at https://github.com/kota200/DEAR-OWL.

Gene set enrichment analysis (GSEA) turns differential expression results into lists of enriched gene sets. These lists are often long and redundant and span several gene-set collections, which makes their biological interpretation difficult. We present lisaR, an R package that reads previously computed differential expression or differential abundance results, runs GSEA and organises the output with the LISA dictionaries (LLM-Inferred Semantic Annotation). The dictionaries place 13,511 gene sets from the three Gene Ontology branches and from curated pathway collections such as Reactome, WikiPathways, KEGG, BioCarta and PID in 158 biological categories grouped into 42 supercategories within four dictionaries. Four open-weight language models scored the semantic fit of each gene set to the categories, and a consensus rule selected the assignments. The dictionaries are distributed with the package and do not depend on the data: studies analysed at different times with the same dictionaries are interpreted with the same categories, and no language model is queried during an analysis. For each category, lisaR summarises the direction of its significant gene sets and applies closed testing with Hommel's robust procedure, which states whether the category shows evidence of enrichment and gives a simultaneous lower bound on how many of its gene sets are enriched. lisaR results are summarised in a report that opens in a web browser and leads from each category to its gene sets and genes, with every figure saved together with its data and R code. We illustrate the workflow with RNA-seq from melanoma biopsies taken before and during nivolumab treatment and with paired tumour and adjacent-tissue proteomics from clear cell renal cell carcinoma. Availability: lisaR is available at https://github.com/DBM-OlmedaLab/lisaR and archived at https://doi.org/10.5281/zenodo.23044930; documentation at https://olmedalab.org/lisaR/.

In network medicine, the protein-protein interaction network, or interactome, is an essential resource for identifying candidate proteins underlying diseases and other phenotypic traits. Indeed, it can be leveraged via the guilt-by-association paradigm, which asserts that interacting proteins are likely to participate in the same molecular processes. Network propagation that combines the topology of the interactome with prior knowledge about disease genes is a promising strategy to identify new candidate genes contributing to diseases. However, existing network propagation algorithms are often biased toward highly connected proteins, called "hubs". We present BFWalk, a novel network propagation algorithm designed to avoid inflated scores for hubs. We tested BFWalk across four human phenotypes: multiple morphological abnormalities of the sperm flagella, dyschromatopsia, hypertrophic cardiomyopathy, and chronic kidney disease. For each phenotype, we assessed whether BFWalk could recover known phenotype-associated genes and whether new candidate genes were enriched in relevant tissues. Depending on the phenotype, BFWalk outperformed or matched the state-of-the-art network propagation methods. Furthermore, the results confirmed that BFWalk is free of bias toward high-degree nodes. Therefore, BFWalk is a powerful method for identifying new genes across diverse phenotypes and constitutes a strong alternative to existing network propagation algorithms. Availability: BFWalk is implemented in Python and C. The code is available under the GNU GPL (v3.0) on: https://github.com/jedrzejkubica/BFWalk

MotivationThe untargeted processing of Liquid Chromatography coupled to High-Resolution Mass Spectrometry (LC-HRMS) data is a major challenge for the comprehensive and robust characterisation of metabolites. In particular, peak detection and annotation are two challenging tasks due to the size and complexity of the data, that are currently addressed independently in existing pipelines, without taking into account the redundancy of information between the ion species of the same compound. Results: We developed an innovative and efficient approach that combines detection and annotation, by 1) representing signals putatively originating from the same molecule as a graph, and 2) inferring the validity of the peaks and their connections within each component using a Graph Attention Network (GAT). We demonstrate on real data sets that the resulting MassGAT model outperforms current approaches in terms of both detection and annotation. Availability and implementation: The MassGAT open-source Python module is publicly available at https://github.com/odisce/MassGAT.

Spatial transcriptomics technologies enable measurement of gene expression while preserving spatial tissue organization, but they remain highly sensitive to technical variability such as library size differences, slide-level effects, and spatial artifacts. Most existing normalization approaches treat normalization as a preprocessing step and perform downstream analyses on normalized values as fixed inputs, ignoring the uncertainty introduced during normalization. We introduce GPSNorm (Gaussian Process Spatial Normalization), a Bayesian spatial normalization framework that jointly models technical variation, spatial structure, and biological signal within a unified hierarchical model. GPSNorm represents gene expression counts using a negative binomial latent Gaussian model whose spatial component is a Gaussian Markov random field approximating a Gaussian process and performs efficient approximate Bayesian inference using the Integrated Nested Laplace Approximation (INLA), producing posterior estimates that propagate normalization uncertainty into downstream differential expression analysis. In simulations anchored to empirical spatial transcriptomics data, GPSNorm accurately recovers spatial technical structure and improves log-fold change estimation compared with existing normalization methods. Applications to three spatial transcriptomics datasets--including human dorsolateral prefrontal cortex Visium data, a GeoMx COVID-19 lung damage study, and the Spatial Organ Atlas kidney dataset--demonstrate improved preservation of biologically expected spatial patterns and marker gene contrasts. These results show that jointly modeling normalization and downstream inference can improve the robustness and interpretability of spatial transcriptomics analyses.An open-source R implementation of GPSNorm is available at https://github.com/Tiny-Quant/GPSNorm.

Background: The three-dimensional organization of the genome is a fundamental aspect of its function and regulation. High-throughput chromosome conformation capture techniques, such as Hi-C, have revolutionized our understanding of spatial genome organization. However, 3C-based methods are resource-intensive and technically demanding. This has driven the development of computational approaches for predicting Hi-C interaction matrices. Hi-cGAN, a novel approach based on conditional generative adversarial networks, offers a computational alternative to extensive wet-lab work by predicting Hi-C interaction matrices. This computational approach contributes to a broader exploration and understanding of genome architecture. Findings: The network pairs a convolutional generator with a convolutional discriminator, evaluated across bin sizes, inputs and cell types. It predicts a whole genome as a cool file at bin sizes from 2 to 25 kb, where Akita, C.Origami and Epiphany emit fixed windows of 1 Mb, 2 Mb and 990 kb. With the input chosen on a validation chromosome, agreement approaches Epiphany's and stays below the sequence-based C.Origami and Akita: over Akita's 411 held-out windows the mean correlation is 0.238 against 0.506. Boundary and loop calls agree less closely, placing the maps at the domain scale. Conclusions: Chromatin factor occupancy determines a substantial part of contact structure, and two tracks capture most of it. The most informative track depends on the resolution: CTCF and the cohesin subunits at 5 to 10 kb, active histone marks at 25 kb. Transfer to an unseen cell type costs about 0.12 SCC, and which method leads depends on the measure.

ABSTRACT Motivation: Myxococcus xanthus is a predatory soil bacterium with a large genome of 9.14 MB due to a genome duplication event. While complete genome sequences of M. xanthus are available, gene annotation remains challenging due to its size and the resulting large number of duplicated genes. Operons, so syntenic block of genes that are co-regulated in bacterial genomes, are an important resource to help predict gene function accurately. Results: In order to help improve the annotation of complex genomes such as the one from M. xanthus, we developed a novel operon prediction tool, DiscOperon, which combines gene expression data with homology searches to identify syntenic blocks: Co-expression data of neighbouring genes across the genome are first used to define gene clusters, which are then used to search for conserved syntenic blocks in fully sequenced bacterial genomes using sequence homology searches. This strategy enables DiscOperon to account for gene insertions, rearrangements and deletions, which is its most distinguishing feature. We have tested DiscOperon against ground truths gene pair information on 3 different species from ODB and RegulonDB and compared it to state-of-the-art and still available operon prediction software and we demonstrate its general usability for operon prediction of any bacterial complete genome. We have applied DiscOperon to predict the operons of M. xanthus, which we are making available for the research community. Availability and implementation: DiscOperon is lightweight, user-friendly python tool with minimal dependencies. It is freely available at https://gitlab.com/habermann_lab/discoperon for general usage. Contact: Theo Brunet (theo.brunet@univ-amu.fr); Bianca Habermann (bianca.habermann@univ-amu.fr). Supplementary information: The operon-structured and annotated M. xanthus genome is available from this manuscript, as well as from Zenodo (https://doi.org/10.5281/zenodo.21976167). We furthermore plan to submit the M. xanthus operon information to the operon database OBD.

MotivationAnalysing and concatenating phage annotation is time-consuming. Further, the output of phage annotation tools cannot be directly submitted to public repositories. To deal with these issues, we developed PhaGAMeToo. This command-line workflow for Linux integrates the functional annotations of two major viral annotation tools (Pharokka and VIBRANT), enabling faster and more accurate functional annotation. Furthermore, the workflow provides merged annotations as submission-ready GenBank files. ResultsPhaGAMeToo uses three steps to generate submission-ready GenBank files. The user uses the reoriented viral genomes as inputs for Pharokka and VIBRANT. Pharokka and VIBRANT-generated files are parsed through the PhaGAMeToo workflow to produce a merged GenBank file. Further, PhaGAMeToo also enables the use of BLASTP to annotate hypothetical proteins not identified by Pharokka and VIBRANT. It then merges the results into a submission-ready GenBank file(s). We tested PhaGAMeToo in three different Use Cases. We analysed reference and uncultivated viral genomes manually curated or directly recovered using MuDoGeR in our Use Cases. In the Use Case 1, we analysed four different NCBI reference genomes. In the Use Cases 2 and 3, we analysed seven recently described huge phage genomes and 56 uncultivated viral genomes recovered from 30 soil metagenomes, respectively. Availability and implementationThe source code, documentation, and installation instructions for PhaGAMeToo are available at https://github.com/NFDI4Microbiota/PhaGAMeToo ContactRene.Kallies@uba.de; ebrardemircioglu25@hacettepe.edu.tr Supplementary informationSupplementary data will be made available upon publication.

MotivationNovel sequencing technologies (for example, different single-cell chemistries and protocols) produce complex data. Often, the sequenced reads themselves encode critical technical information, such as the cell or molecule of origin. For effective pre-processing of this data and subsequent downstream analysis, it is required to efficiently and accurately identify, extract, and potentially normalize this information. ResultsWe introduce seqproc, a general-purpose sequence pre-processing tool based on a concise descriptive grammar to specify sequence matching and transformations. seqproc compiles a user-provided sequence geometry and transformation description into an execution graph, executed by the antisequence library. We demonstrate that seqproc is faster on most chemistries, substantially more memory efficient, and at least as accurate as alternative tools that provide similar functionality, while having a more concise description syntax. Availabilityseqproc is written in Rust and can be executed as a binary program or used as a Rust crate. It is licensed under the BSD 3-clause license and the source code is available at https://github.com/COMBINE-lab/seqprocy.

Background: Weighted Gene Co-expression Network Analysis (WGCNA) is a widely adopted systems biology method to discover gene modules and module-trait associations, mostly from transcriptomics. Designed for a single layer, it cannot jointly analyze multi-omics layers, a consequential limitation in modern biomedical research. WGCNA modules are often hard to interpret, requiring vast follow-up for contextualization. Moreover, no integrated framework exists to visualize condition-specific, cross-omics relationships at module or feature level. Results: To address these limitations, we developed WGCNA+, a novel R package extending WGCNA to multi-omics. WGCNA+ offers key innovations: (i) a unified multi-omics pipeline for per-layer network inference and cross-layer module enrichment; (ii) SVD-accelerated topological overlap matrix calculation that greatly reduces computation time; (iii) a consensus framework identifying modules reproducible across independent datasets/conditions; (iv) LASAGNA, a companion R package for phenotype-conditioned, multi-partite graph visualization of cross-omics relationships; (v) AI-powered annotation and infographics offering immediate biological insight. We tested WGCNA+ across public transcriptomics, proteomics, and miRNA datasets. WGCNA+ detects biologically meaningful modules, cross-omics feature and phenotype correlations, and provides AI-powered interpretation that accelerates research. Conclusions: WGCNA+ addresses existing gaps with a principled, efficient framework for co-expression network analysis across omics. It detects cross-omics regulatory modules and their phenotype association to support basic research, biomarker discovery and pathway analysis. It uniquely offers AI-assisted interpretation and infographics, aiding hypothesis generation. Complementing WGCNA+, LASAGNA is a phenotype-aware multi-partite visualization framework to explore cross-omics relationships. Altogether, these features make WGCNA+ an innovative, powerful tool for clinical and translational research. Availability and implementation: WGCNA+ and LASAGNA are implemented in R language for statistical computing, version[&ge;]3.5. WGCNA+ and LASAGNA are fully and freely available with no restrictions (https://github.com/bigomics/WGCNAplus; https://github.com/bigomics/lasagna)

Summarybcftools is the standard toolkit for handling VCF and BCF variant files, but it processes records on a single core; its --threads option speeds up only compression of the output, not the work done on variant records. Processing large call sets is therefore slow, and users often divide the genome and reassemble the results by hand. We present pbcftools, a Perl wrapper that does this automatically: it splits the genome into chunks, runs an ordinary bcftools command on each in parallel, and reassembles the outputs by a method suited to the data type. Across Linux servers, Windows/WSL2 workstations and Apple laptops, with bcftools 1.21 to 1.24, parallel output was identical to serial output for every command tested. On 1000 Genomes Phase 3 data, operations writing compressed VCF ran 10.8 to 21.1 times faster with 32 cores and up to 35.4 times with 64, those writing text 3.7 to 12.8 times, and merging 100 VCF files 19.2 times. pbcftools also runs on LSF and Slurm clusters. Availability and implementationpbcftools is written in Perl (>= 5.16) and requires bcftools; local parallel execution also requires Perl module Parallel::ForkManager. It is released under the MIT license at https://github.com/zhangge-uc/pbcftools (DOI: 10.5281/zenodo.21780361).

Motivation: Biomedical natural language processing (Bio-NLP) classification tasks are often limited by the cost of manual annotation, especially for specialised extraction problems where labelled corpora are scarce. Active learning can reduce this cost by prioritising informative samples, while large language models (LLMs) avoid task-specific annotation, but at high computational cost. We ask whether an encoder trained via active learning can match LLM performance at a lower computational cost, using identification of chondrogenesis regulators from PubMed abstracts as a test case applicable to other under-annotated Bio-NLP problems. Results: We built an active learning framework to fine-tune PubMedBERT to classify gene mentions as regulators of chondrocyte differentiation, achieving an AUC-ROC of 0.93 using 1,238 annotated sentences. Against open-weight LLMs (Qwen3, Llama-3.1), PubMedBERT matched the top-performing LLMs on AUC-ROC and achieved the highest precision (0.69) of any model, while classifying the full corpus far faster. Applied to 67,111 gene mentions, it identified 1,128 candidate regulators, recovered 79% of Gene Ontology-annotated genes related to chondrocyte differentiation (94 genes recovered), and proposed 981 new candidates. The pipeline is agnostic to the target process, suggesting it could be adapted to identify regulators of other biological processes. Availability and Implementation: Code is available at https://github.com/ChondroTextomics/ALRegulatorDiscovery. Data are available at Zenodo: https://doi.org/10.5281/zenodo.22746090. Final model is available in HuggingFace Hub amav/pubmedbert-chondrogenesis-classifier Supplementary information: Supplementary data are available in a separate file