Discover, rate, and discuss biology preprints.

The paper introduces CCIDeconv, a method to deconvolve whole-cell cell-cell interactions into subcellular compartments using scRNA-seq data and spatial transcriptomics datasets.

This paper develops a neural network model, nnCNV, to detect CYP2C19 deletions from genotyping array signals, achieving higher accuracy than existing methods.

MPGEM harmonizes and completes gene-expression profiles from legacy microarray datasets, enabling large-scale transcriptomic analysis.

The paper develops a single-cell transcriptomic reference atlas and evaluates a transcription factor recipe for reprogramming fibroblasts into hematopoietic stem cells, revealing intermediate transcriptomic states and transcriptional remodeling.

The paper introduces three R/bioconductor packages to address technical variations in ATAC-seq data and improve motif accessibility analysis.

MultiFlow predicts single-cell multiomic perturbation responses by jointly modeling gene expression and chromatin accessibility across cellular contexts.

This paper introduces a statistical framework for disease classification using scRNA-seq data, achieving better interpretability and performance compared to existing methods.

Trimole-Hybrid improves ADMET prediction by combining multiple molecular representations and task-specific models, achieving top performance on 22 benchmarks.

OMIO provides a policy-driven Python library to standardize microscopy image I/O, ensuring reproducible and interoperable data handling across formats and workflows.

HuMMANet harmonizes gut microbiome and metabolome data across 46 studies to identify reproducible metabolite-microbiome associations and their clinical relevance.

GTX-GUT is a standardized metagenomic workflow for gut microbiome profiling that integrates quality control, taxonomic classification, and clinical association analysis, showing improved accuracy and reproducibility in microbiome studies.

The paper identifies a subset of proteins that causes discordance between structure and disorder predictors, showing that this subset is compositionally distinct and not a general sequence-class artefact.

The paper introduces a dynamic alert system using a Zoonotic Alert Index to improve disease outbreak detection in wildlife, validated with yellow fever data achieving over 94% accuracy.

This paper presents a pipeline for polygenic risk score calculation that includes data curation, statistical methods, and reporting, resulting in a curated database and interpretable reports for PRS analysis.

This study compares the effects of incomplete lineage sorting and gene tree estimation error on species tree inference using simulated and real datasets.

This study benchmarks segmentation models' robustness to corruptions in biological imaging, finding that older methods can be more robust than newer foundation models.

PhageLysData is a curated, AI-ready dataset of phage lytic enzymes and depolymerases, providing annotated sequences, structural data, and embeddings for machine learning applications.

BIOCURRENT infers donor-specific pseudotime geometry in transcriptomic data by modeling gene expression with microenvironmental context and latent pseudotime, revealing condition- and donor-specific distortions in progression intervals.

The paper evaluates missense pathogenicity predictors using ClinVar data, finding that gene identity, not variant effect, largely drives benchmark performance.

CodonMamba is a foundation model for mRNA coding sequence design that achieves state-of-the-art performance and enables programmable codon optimization through inference-time steering.