Discover, rate, and discuss biology preprints.

The paper uses AlphaGenome to assess the transcriptional impact of somatic mutations in aging, finding their effects are much smaller than endogenous aging transcriptional changes.

CLEAR-HPV discovers interpretable morphologic concepts in HPV-associated histology slides using attention-based methods, reducing high-dimensional features to 10 concepts while preserving predictive information.

The paper develops simCRISPR, a simulation framework for pooled CRISPR screens, evaluates normalization strategies for detecting gene-by-environment interactions, and shows improved results with safe-harbor normalization.

This study uses in-silico methods to investigate the interaction between 20-HETE and NKA, identifying key amino acids involved in ligand-receptor binding.

This paper improves variant effect prediction by modifying protein language models with sparse mechanistic features, achieving significant performance gains across multiple datasets.

PromoterAtlas uses a transformer model to predict regulatory sequences in Gammaproteobacteria, showing cross-species recognition of promoters and sigma factors.

PlantP450Dock automates molecular docking for plant P450s by integrating heme cofactor implantation, conformational sampling, and data-driven residue selection to improve docking accuracy.

APEX identifies pathogen targets and designs inhibitors using explainable AI, highlighting key residues and guiding molecular design.

This study shows that deep learning models for chemical perturbation prediction do not effectively use drug molecular features, as performance remains unchanged when drug inputs are removed.

The paper analyzes the conformational landscapes of frameshifting elements (FSEs) across multiple viruses, revealing length-dependent plasticity and antisense-driven structural changes.

This paper introduces a metabolic model for large language models (LLMs) that imposes energy constraints to simulate existential vulnerability and emergent agency.

The paper introduces a physics-informed autoencoder for anomaly detection in GBM RNA-seq data, achieving higher AUC than standard methods.

This paper introduces a cross-domain spatial transcriptomic framework to analyze tissue repair dynamics, showing coherent tissue state separation and distinct spatial organization patterns in wound healing versus tumor microenvironments.

Ensemblify generates and analyzes conformational ensembles of intrinsically disordered proteins using Monte Carlo sampling and AlphaFold confidence metrics.

The paper identifies a conserved helical motif in DCAF7 partners using structural and computational methods, revealing its role in scaffolding and linking to neurodevelopmental disorders.

LAMPrEY is a Python-based quality control tool for large-scale proteomics datasets that automates processing and analysis with MaxQuant and RawTools, enabling reproducible and rapid data interpretation.

The paper develops TriCyP, a model to predict cysteine states in protein structures, achieving high accuracy and identifying new biological patterns and protein families.

PL-PatchSurfer3 improves structure-based virtual screening by incorporating hydrogen bond complementarity and visibility features, enhancing performance while maintaining tolerance to receptor structure changes.

MethylCurate automates DNA methylation dataset retrieval and metadata harmonization from Gene Expression Omnibus, enabling scalable evaluation of epigenetic aging clocks.

The paper introduces Aiki-Sol, a predictor for E. coli protein solubility that accounts for centrifugation regimes, improving performance by treating regime as a feature rather than noise.