bioRxiv Science⌕ Search

Biology subjects

Zandigohar, M.

Publications and source records attributed to Zandigohar, M..

2 recordsLinked to original sources

RAGulate: Retrieval-Augmented Generation for Post-hoc Literature-Grounded Regulatory Assessment

MotivationPrioritization of transcription factor (TF)-target relationships predicted by computational models for experimental validation often requires biologists to manually inspect heterogeneous and context-dependent evidence scattered across the biomedical literature. Large Language Models (LLMs) offer a promising solution to streamline this task. However, their reliance on general-purpose knowledge may lead to hallucinations and inaccurate interpretations. ResultsWe present RAGulate, a retrieval-augmented generation (RAG) framework for literature-grounded assessment of transcriptional regulation. RAGulate leverages CollecTRI, an external regulatory knowledge base, and integrates alias-aware query expansion, sparse and dense retrieval, maximum-marginal-relevance re-ranking, and LLM-based classification of predictions within a modular pipeline. Using a balanced TF-target-context benchmark from the same resource, we evaluate retrieval, classification, and evidence faithfulness. While CollecTRI provides TF-target links with supporting PubMed Identifiers (PMIDs), RAGulate infers the context of each interaction from the retrieved literature. Results show that alias normalization markedly improves retrieval recall, while hybrid retrieval, which merges lexical and embedding-based candidates, achieves the highest evidence recovery across all cut-offs. Conditioning LLMs on retrieved documents consistently improves AUROC and AUPR for classifying whether a TF-target interaction is supported in the specified context compared with direct prompting. RAGulate reduces hallucinations and improves PMID-level citation correctness, producing explanations that faithfully reflect the supporting literature. RAGulate represents a knowledge-based AI tool that partners with biologists to accelerate the process of TF-target prioritization for experimental validation and foster hypothesis generation. Availability and implementationThe software and tutorials are available at github.com/YDaiLab/RAGulate.

bioinformatics↗

scRegulate: Single-Cell Regulatory-Embedded Variational Inference of Transcription Factor Activity from Gene Expression

MotivationAccurately inferring transcription factor (TF) activity from single-cell RNA sequencing (scRNA-seq) data remains a fundamental challenge in computational biology. While existing methods rely on statistical models, motif enrichment, or prior-based inference, they often depend on deterministic assumptions about regulatory relationships and rely on static regulatory databases. Few approaches effectively integrate prior biological knowledge with data-driven inference to capture novel, dynamic, and context-specific regulatory interactions. ResultsTo address these limitations, we develop scRegulate, a generative deep learning framework leveraging variational inference to estimate TF activities guided by experimental TF-target gene relationships and progressively adapted based on the input scRNA-seq data. By integrating structured biological constraints with a probabilistic latent space model, scRegulate offers a scalable and biologically grounded estimation of TF activity and gene regulatory network (GRN). Comprehensively bench-marking on public experimental and synthetic datasets demonstrates scRegulates superior ability. Further, scRegulate accurately recapitulates experimentally validated TF knockdown effects on a Perturb-seq dataset for key TFs. Applied to experimental human PBMC scRNA-seq data, scRegulate infers cell-type-specific GRNs and identifies differentially active TFs aligned with known regulatory pathways. scRegulates TF activity representations capture transcriptional heterogeneity, enabling accurate clustering of cell types. scRegulate is highly efficient, frequently an order of magnitude faster than common baselines. Collectively, our results establish scRegulate as a powerful, interpretable, and scalable framework for inferring TF activities and GRNs from single-cell transcriptomics. AvailabilityResults and scripts available at github.com/YDaiLab/scRegulate. Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗