Discover, rate, and discuss biology preprints.

This study uses deep learning to predict neoadjuvant chemotherapy benefit in breast cancer using stromal histology, introducing the TS-score which outperforms existing predictors.

This study identifies epsin-2 as a therapeutic target for multiple system atrophy by regulating FABP7/-synuclein hetero-aggregates propagation.

This study identifies PnPf2 as a central transcriptional regulator of virulence in Parastagonospora nodorum by characterizing its genomic targets and associated transcription factors.

This paper develops an integrative deep learning model combining image, clinical, and genomic data to improve risk stratification for colon adenocarcinoma patients, showing better performance than models using single data types.

This study investigates how Mikania micrantha responds to Puccinia spegazzinii infection using metabolomics and transcriptomics, identifying changes in metabolites and gene expression.

This study compares transcriptome profiles of trypanotolerant and trypanosusceptible cattle breeds during Trypanosoma congolense infection, revealing differences in immune response and metabolic regulation.

This paper develops a CRISPR-based point-of-care test for detecting ToBRFV, showing specificity and comparable sensitivity to PCR, with a simplified nucleic acid extraction protocol for field use.

This study shows that persistent macrophages after AKI promote a proinflammatory environment leading to tubule damage and CKD progression.

This study shows that higher active TGFβ1 levels in human subchondral bone correlate with cartilage loss and poor bone quality in knee OA.

scoMorphoFISH provides a deep-learning tool for spatial single-cell mRNA quantification and morphometry in kidney biopsies, enabling single-mRNA resolution analysis.

This study analyzes global salivary gland tumor data from 15 centers to provide updated epidemiological insights on incidence, demographics, and tumor types.

This paper develops an automated machine learning pipeline for differential cell counts on whole-slide bone marrow aspirate smears, showing high concordance with manual counts and reduced intra-slide variance.

This study shows that inhibiting ATR in C. elegans reduces the severity of mitochondrial ETC dysfunction by reversing nucleotide imbalances and restoring translation.

The paper develops SegPath, a large annotation dataset for histopathology segmentation, showing it outperforms pathologist annotations and reduces bias.

This study shows that hypereosinophilia in mice leads to cardiac inflammation and impaired cardiomyocyte contractility, contributing to heart dysfunction.

This study uses digital spatial profiling to analyze transcriptional heterogeneity in collapsing glomerulopathy lesions from HIV and SARS-CoV-2 infected patients.

This paper develops a deep learning-based preprocessing tool to analyze urothelial cell clusters in urine cytology, identifying key markers of atypia that could improve machine learning applications in bladder cancer detection.

This study evaluates four mosquito collection methods for xenomonitoring lymphatic filariasis and malaria in Burkina Faso, finding Human Landing Catches (HLC) most effective for collecting mosquitoes and Window Exit-Trap and Pyrethrum Spray Catches (PSC) effective for sampling infected vectors.

This paper analyzes whether deep learning models in histopathology learn trustworthy or deceptive strategies, finding that some recent advances in molecular profiling from H&E slides are unreliable.

This study compares high- and low-level deep learning approaches for microsatellite instability prediction using TCGA data, finding the high-level approach superior and identifying key morphological features.