Discover, rate, and verify biology preprints.

This study evaluates LLMs' ability to predict mutation effects based on structural data without using dedicated predictors, finding comparable accuracy to existing methods.

This paper evaluates the performance of intermediate layers in single-cell foundation models for cell-type and perturbation prediction, finding that some layers perform as well as final layers.

The paper benchmarks single-cell foundation models for cell-type and disease-state prediction, finding they outperform baselines in low-data regimes.

This paper introduces HARMONIC, a framework that improves cell-cell communication inference in spatial transcriptomics by incorporating histology data, showing it better captures anatomical context.

AlphaFold3 pLDDT matches AlphaFold2 at calling disordered residues (Matthews correlation 0.693 vs 0.701); on disorder-containing complexes mean DockQ is similar, but interface features explain less variance (42% vs 58%).

The paper develops a Bayesian gene interaction network to identify 13 genes for PD classification, achieving high AUC in a blood-based transcriptomic analysis of the PPMI cohort.

Tree models on RDKit/Morgan features predict M.tb activity and MIC endpoints for drug-susceptible and resistant isolates, slightly beating pretrained embeddings; released as a Colab tool.

This study shows that selecting optimal bin sizes improves tumor cell phylogeny inference from low-coverage single-cell sequencing data.

The paper shows that horizontal gene transfers can reveal information about extinct and unsampled lineages in species trees through gene tree analysis.

The paper designs novel inhibitors targeting thiamine biosynthesis in bacteria using structure-based drug design, resulting in a potent orally bioavailable broad-spectrum bactericidal molecule.

The paper accelerates structural analysis in HADDOCK3 by reformulating RMSD and FCC clustering as GPU-based linear algebra operations, achieving significant speedups without compromising accuracy.

The paper uses GenePy scores and clustering to identify genetic clusters in IBD patients, finding NOD2 and NLRP1 as significant drivers with phenotypic associations.

This paper shows that biological concepts can be represented as directions in foundation model embeddings, enabling interpretable and reusable analysis of diverse biological properties.

The paper introduces crispAIBE, a probabilistic framework for predicting base editing efficiency and bystander allele composition, achieving high correlation and coverage on benchmark datasets.

This paper introduces NanoPCT, a framework that predicts protein corona formation by learning universal patterns across proteins and enables zero-shot prediction for unseen proteins.

KEDROS is a web platform that performs phylogenetic profiling across proteomes to identify lineage-specific genes and functional divergence, offering a user-defined workflow for comparative genomics.

This paper introduces zPolaron, a graph neural network for functional protein design that integrates structural, interaction, and evolutionary information, improving fitness prediction beyond structural compatibility.

The paper evaluates models for classifying synonymous variants by separating gene context, exon position, and variant-specific effects, finding that gene identity and position are more predictive than the variant itself.

SORT identifies multilevel gene programs and their spatial activities across diverse spatial transcriptomic datasets, revealing both shared and sample-specific patterns.