Discover, rate, and discuss biology preprints.

MOSANIC learns a spatial cell-cell communication network from transcriptomics data to identify network hubs and predict vulnerabilities in breast cancer.

This paper introduces a model-guided virtual screening pipeline to improve synthesizability in generative drug design by combining SBDD models with chemical similarity search.

PRCGAP provides a haplotype-resolved analysis of cancer genomes, detecting variants missed by standard references and revealing new insights into structural variants and fusion events in leukemia.

The paper introduces gecco, a gene-specific normalization method for bulk RNA-seq that reduces library size bias while preserving biological signal, improving differential expression and pathway analysis.

Rectangle provides a scalable deconvolution framework for bulk RNA-seq using single-cell data, achieving high accuracy and robustness across diverse benchmarks.

NicheTrans translates single-omics data into spatial multi-omics insights by integrating spatial and multi-modal information, outperforming existing methods in disease and tissue analysis.

This paper introduces coding agents that generate interpretable single-cell embeddings from literature-based gene programs without training data or databases.

IgGM2 is an all-atom foundation model for immune receptor design that jointly predicts CDR sequences and full-atom structures, outperforming existing methods in structure prediction and sequence design benchmarks.

The paper introduces a benchmark for scientific coding agents in spatial transcriptomics alignment, showing that richer context improves tool exploration but reduces alignment accuracy.

BayesInteractomics improves AP-MS interactome analysis by combining deep learning and Bayesian statistics to better distinguish true interactions from background noise.

GeneLLM uses language models to detect cancer from cfRNA sequences, identifying novel biomarkers from unannotated regions with improved accuracy.

This paper shows that general-purpose language models can detect protein homology without biological pre-training, achieving ROC-AUC 0.79 on a controlled benchmark.

VLab4Mic predicts structural resolvability in super-resolution microscopy by simulating imaging configurations and comparing results to experimental data.

BBBP-Atlas predicts blood-brain barrier permeability for small molecules and peptides using a cross-modal dataset and structure-aware model, achieving high accuracy and outperforming existing methods.

The paper introduces metrics for assessing tensor decomposition quality in immune profiling data using a dictionary learning framework.

The paper quantifies how mutational biases shape protein evolution in RNA viruses by analyzing SARS-CoV-2 and 34 other species, showing that mutational spectra predict amino acid composition changes and deviations indicate selection pressures.

This paper introduces Gene Program Negotiation (GPN), a framework to analyze regulatory interactions in single-cell transcriptomes, revealing new regulatory states and their association with cellular plasticity.

DyViA uses generative AI to create personalized aging trajectory models for femoral neck BMD, showing dynamic progression from 66 to 89 years.

qg generates reproducible, configuration-driven acquisition queues for mass spectrometry with QC control and vendor-specific file output.

This paper introduces stGP, a statistical framework to identify cell-type-specific spatiotemporal gene programs, revealing how gene programs are influenced by time and spatial structure in tissues.