Discover, rate, and discuss biology preprints.

ChatSpatial enables reproducible spatial transcriptomics analysis by using an LLM to select and orchestrate tools with schema enforcement, improving parameter validity and workflow consistency.

The paper evaluates DNA foundation models' ability to predict R-loop-forming sequences, finding they perform well in some settings but struggle with cross-platform and cross-species generalization.

The paper uses language modeling to create a comprehensive representation of protein biology, enabling structure prediction and discovery of high-affinity protein interactions.

OmniGene-4 is a multi-modal bio-language MoE model that decomposes task performance by router states, achieving strong benchmarks in homology and biological-knowledge tasks while being computationally efficient.

SciCore-Omics is a tri-modal foundation model that integrates histology, spatial transcriptomics, and language to improve spatial biology analysis and interpretation.

The paper analyzes genomic and transcriptomic data from two Tunisian durum wheat landraces to identify adaptive mechanisms under climate stress.

This paper designs and validates a high-affinity peptide inhibitor targeting the HPV E1-E2 interface using computational methods, showing improved binding affinity and safety.

UnBlender enables cell type deconvolution in respiratory bulk RNA-seq data and provides routine validation of deconvolution accuracy.

This study evaluates a pretrained graph neural network for predicting P-glycoprotein substrate status using a benchmark dataset, achieving high AUROC scores.

TissueFormer predicts population-level phenotypes from single-cell data, outperforming existing methods in predicting COVID-19 severity and cortical area identity.

This paper identifies hypoxia-related genes associated with prognosis in lung adenocarcinoma, finding DSG2, EIF6, and EXO1 as independent prognostic factors.

This study identifies protein domains under positive selection in tumor genomes, revealing convergent evolution in core cellular processes and regulatory functions.

This study compares age estimation models using DNA methylation data, finding that maximum likelihood estimation reduces bias compared to linear regression and random forest.

SpatioCell integrates histology and spatial transcriptomics to extract single-cell resolution cell type and expression information from ST data, revealing microenvironment features in triple negative breast cancer.

This paper develops AdventML, a model for predicting enzyme optimal temperature using transformer embeddings and resampling, achieving R2 ~0.64 and improved extreme temperature predictions.

AI-IDP predicts conformational ensembles of disordered proteins by combining deep learning with physical restraints, revealing common structural features across proteomes.

This study evaluates and improves low-frequency somatic mutation detection across multiple sequencing platforms using a simulated benchmark and a machine learning-based caller.

CodeCytos improves spatial molecular imaging analysis by enabling dynamic, programmable interaction with tissue data, outperforming baseline approaches in custom feature exploration.

CellClick is an interactive platform that improves cell type annotation accuracy in single-cell and spatial omics data through automated and user-guided tools.

QMAP evaluates antimicrobial peptide MIC and hemolytic activity prediction models using standardized, homology-constrained benchmarks, revealing limited progress and poor performance in high-potency predictions.