Discover, rate, and discuss biology papers.

The paper introduces pre-miRBench, a multispecies benchmark for pre-miRNA prediction, showing its model outperforms existing predictors in cross-species and cross-family tests.

PREpiBind evaluates ten protein representations for pMHC-II binding prediction across multiple datasets, showing varying performance depending on the task and evaluation method.

The paper develops likelihood ratios for forensic DNA mixtures from shotgun sequencing data, showing they can be unreliable and proposing a combined evidence score to improve accuracy.

PanRes is a curated database of antimicrobial resistance genes that uses structural clustering and HMMs to improve detection of resistance determinants in metagenomes.

The paper develops a proteomic aging clock using longitudinal plasma proteomics data to show that biological aging is dynamically influenced by modifiable risk factors.

HiFIseek identifies gene-specific enrichment of high-impact mutations in genomic regions associated with cancer and breast cancer risk SNPs, showing limited agreement across CRR-score combinations.

The paper characterizes cross-family conserved fragments, showing their structural and biophysical properties and their role in folding and stability.

The paper introduces digestome, a marker-gene panel for functional profiling of anaerobic digestion microbiomes, and benchmarks it against metagenome-assembled genomes and species representatives.

RGI-Toolkit enables controllable biomolecular structure prediction by integrating experimental restraints into diffusion models without retraining.

The paper evaluates an agentic system for drug discovery, finding that its performance depends on predictive models rather than generation methods, and that outputs are indistinguishable from expert work.

The paper introduces Flock, a large negative-enriched PPI dataset to improve computational prediction methods by providing high-quality negative interactions and a benchmark for evaluating models.

Predicting differentially expressed genes and expression shifts per condition from database embeddings plus per-gene evidence, PIE reaches 1.2–3.2 times the best baseline's AUPRC on unseen contexts, perturbations and datasets.

The paper derives a closed-form cell quota to determine the number of cells needed to accurately estimate perturbation direction in single-cell screens, validated across multiple datasets.

This study examines verification debt in agentic AI workflows for biological research, using Biomni to analyze 10 test cases across proteomics, genomics, and drug discovery.

This study identifies backdoor vulnerabilities in genomic foundation models through data poisoning during fine-tuning and proposes an auditing framework to detect such threats.

GARAGE generates realistic synthetic scRNA-seq data by addressing high-dimensional, small-sample and rare-cell challenges using graph attention and adversarial generation.

The paper uses shotgun metagenomic sequencing to compare bacterial communities in plaque, saliva, and tumor samples from OSCC patients and healthy controls, identifying distinct microbial profiles.

This paper develops a machine learning framework to predict antibacterial synergy using multi-level bioactivity features and demonstrates its performance under rigorous cross-validation.

The paper introduces an R package for EndoMT gene set enrichment analysis, showing its utility in detecting EndoMT-associated transcriptional changes across different experimental conditions.

methylTFR quantifies transcription factor activity from DNA methylation, identifying cell-type-specific regulators and resolving developmental trajectories in immune cells.