Discover, rate, and discuss biology preprints.

This paper introduces STAMP, a machine learning framework for predicting antimicrobial peptide efficacy across bacterial species, achieving strong predictive performance and real-world validation.

AEGIS standardizes and validates genome annotation files, enabling reliable downstream analysis through error correction and comparative genomics tools.

The paper presents results from an open competition to develop aging biomarkers using DNA methylation data, showing improved predictive models for age, mortality, and multi-morbidity.

Aardvark benchmarks genomic variant calling accuracy by comparing haplotype sequences with a basepair score, supporting multiple variant types and outperforming hap.py.

This paper introduces Wasserstein Geodesic Transformations to improve analysis of high-throughput cell imaging data by addressing long-tailed distributions and outliers.

This paper introduces a novel method for efficiently generating yield spaces from genome-scale metabolic networks, enabling advanced analysis techniques for metabolic modeling.

LinearCapR computes per-nucleotide structural-context probabilities for RNA in linear time without span limits, outperforming existing methods on long-range structures and large RNAs.

This study evaluates six epistasis detection methods for quantitative phenotypes using simulated data and the ABCD dataset, finding varying performance across interaction types.

ALARMIST identifies multicellular communication motifs in spatial transcriptomics data, revealing microenvironmental drivers of tumor progression in lung and brain cancers.

The paper identifies 80 RNA virus species from molluscan transcriptomes, revealing extensive diversity, modular genome evolution, and host interactions.

The paper investigates methylation patterns of polymorphic transposable elements using long-read sequencing and the Human Pangenome Reference, revealing distinct methylation profiles and context-dependent interactions.

barbieQ is an R package for analysing barcode count data from clonal tracking experiments, providing statistical tests for differential barcode proportion and occurrence.

This paper develops data-driven sampling strategies to improve bird detection models by selecting the most impactful samples for fine-tuning with limited annotations.

TESSERA is a foundation model for the cancer genome that uses self-supervised learning to predict variant pathogenicity, tumour typing, and treatment response, revealing interpretable biomarkers for chemotherapy selection.

This study benchmarks seven variant callers for detecting low-frequency variants in M. tuberculosis using simulated short-read data, finding FreeBayes performs best and introducing an error model to filter false positives.

Monju performs multi-criteria clustering on single-cell omics data by integrating multiple clustering solutions to assess cluster stability.

The paper identifies phytochemicals with multi-target activity in psoriasis by integrating transcriptomic and systems biology analyses to reveal novel regulatory mechanisms and drug candidates.

BioMedGraphica integrates biomedical knowledge into a unified graph and introduces a Textual-Numeric Graph structure for multi-omics analysis.

StrainCascade automates bacterial genome reconstruction and characterization using long-read sequencing, enabling high-resolution comparative genomics and functional analysis.