Discover, rate, and discuss biology preprints.

SPaRTA enables non-experts to query spatial-omics data via natural language, achieving 63-77% accuracy with fine-tuning on cancer tissue datasets.

The paper evaluates an eight-gene module associated with trogocytosis in CRC, finding its correlation with T-cell fraction is typical of random modules, suggesting it reflects cellular composition rather than trogocytosis.

This paper introduces LSAT-EN, a transfer learning method that improves survival prediction for younger breast cancer patients using multi-omics data from older patients.

The paper introduces a framework to analyze pan-cancer co-expression programs by addressing artifacts in existing methods, showing that gene expression patterns differ between cancer types.

The paper introduces AtomLM-9b, a language model using an all-atom structure tokenizer to predict protein structures with performance comparable to AlphaFold 3.

PPI-Caliper controls false discovery rate in protein interaction prediction using conformal calibration without retraining models.

PLATE provides a global resource of food metaplasmidomes, offering curated catalogs and annotations for studying plasmid diversity and function in food ecosystems.

This paper introduces AutoSPECTRA, a protocol for auditing generalizability of scientific foundation models across biological domains.

This paper introduces XAMPGen, an interpretable generative model for discovering novel antimicrobial peptides using fine-tuned protein language models and explainable AI techniques.

This paper develops a 2D U-NET model to segment and quantify atherosclerotic lesions in mouse aortic sections, achieving high accuracy and significantly reducing measurement time.

This paper introduces a model-agnostic decomposition method to analyze differences in mean response between groups, using game-theoretic attribution and optimal transport coupling.

The paper shows that neural RNA splicing models rely on reading frame information that the splicing machinery cannot access, leading to errors in predicting poison exons.

RAPTORGraph improves causal discovery in single-cell perturbations by modeling gene interactions and isolating perturbation effects, outperforming existing methods on downstream analyses.

GOTSurv models patient-specific disease progression from cross-sectional survival data, identifying heterogeneous risk strata and individualized progression trajectories without predefined risk cut-offs.

Peptorama provides tools for creating and analyzing non-ribosomal peptide structures, enabling efficient comparison of known and predicted peptides to aid in biodiscovery.

The paper analyzes drug response direction across glioma states using single-cell RNA-seq data, finding that response direction is consistent across states but magnitude varies.

Rho is a collaborative analysis environment that organizes materials by scientific purpose to support human-AI judgment in bioinformatics.

CARTnet improves B-cell epitope prediction using a contact-aware deep learning framework, outperforming existing methods on internal and external datasets.

The paper compares RNA sequencing protocols using long-read sequencing, finding significant changes in isoform structure and abundance due to RNA processing steps.

This study introduces QUEST, a model that uses H&E histology to predict spatial protein expression, enabling virtual staining and analysis across biological scales.