Discover, rate, and discuss biology preprints.

The paper shows label noise limits TCR-pMHC specificity prediction and improves performance using AlphaFold3 and denoising.

OmicsFM is a proteomics foundation model that matches transcriptomic models in performance, revealing complementary biology between proteomics and transcriptomics.

This paper develops a deep learning framework to infer gene expression from cell-free DNA, improving reliability and enabling functional genomics in precision oncology.

FigTreeKit provides a Python toolkit for programmatically styling and auditing phylogenetic trees, with scalability testing on large bacterial and archaeal datasets.

The paper evaluates how different model architectures, LD filtering thresholds, and variant prioritization methods affect gene-level eQTL score prediction quality using over 3 million models.

This study surveys public awareness of the gut-brain axis and compares microbial dysbiosis in depression, anxiety, schizophrenia, and ASD using the Disbiome database.

InsilicoCell predicts cellular functional profiles and enables multi-objective virtual drug screening by integrating diverse molecular data and tasks.

The paper introduces morphOTU, a framework using phenotypic data to create operational biodiversity units, achieving diversity estimates comparable to expert identifications.

PMPNN-DDG is a Random Forest model predicting single-point mutation-induced protein stability changes using features from ProteinMPNN, achieving high accuracy on independent test sets.

DeMoP uses a language-model-guided mixture-of-experts framework to improve cancer prognosis by integrating clinical and molecular data across cohorts.

This paper introduces MT-LLE, a multi-task manifold learning method that improves disease modeling from longitudinal omics data by integrating clinical labels and temporal progression, achieving better clinical structure and progression forecasting than standard methods.

ASAREE evaluates agentic AI system performance and resource efficiency through factorial experiments, showing that advanced models and critic agents increase costs without improving predictive performance.

ROADIES-XP enables efficient incremental species tree updates from raw genomic assemblies using GPU acceleration, achieving high speedups and maintaining topological accuracy.

DeepTMHMM2 predicts transmembrane protein topology and subcellular location, including re-entrant regions and interfacial helices, and identifies localization across 17 membranes.

HIDE-Deconv improves deconvolution by analyzing cellular composition across multiple resolution levels, revealing hidden cellular remodeling in diseases.

TaHL-PTM improves PTM prediction by addressing tokenization issues in decoder-only PLMs, achieving better performance on residue-level tasks.

This study identifies novel HDAC11 inhibitors through in silico and in vitro approaches, showing they reduce breast cancer cell viability and induce apoptosis.

This paper introduces RCDO, a closed-loop reinforcement learning framework that accelerates compound optimization by integrating experimental feedback with generative models, achieving significant improvements in drug candidate properties.

The paper benchmarks imputation methods on scRNA-seq data from AMI patients to assess their impact on downstream analysis.

The paper identifies age-associated changes in the skeletal muscle epigenome and transcriptome, revealing shifts in chromatin accessibility and gene expression linked to metabolic and senescence pathways.