Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Single-cell
Learned models map single-cell and spatial omics to cell states and regulatory elements
Recent preprints and publications couple representation learning with generative, multimodal and spatial readouts across single-cell modalities.
Tools & infrastructure
New tools for metagenomic binning, pathogen detection, and fungal ITS profiling
Packaged, benchmarked tools for metagenomic, amplicon, and proteomic analysis appeared as preprints and publications dated April 2026.
ML / AI methods
Learned sequence embeddings with graph networks predict biomolecular interactions
Pretrained ESM and genomic sequence models are increasingly fused with structure and graph encoders across interaction and regulation tasks.
Epigenomics
Baseline modeling recovers perturbation signatures from pooled single-cell screens
Recent work models baselines, nested structure, and RNA-protein discordance explicitly, as confounds limit interpretation of molecular profiles.
Structural bio & protein design
Learned and knowledge-based models score protein-ligand complexes for affinity and pose
Graph networks and direction-aware potentials now separate binding affinity from pose prediction, benchmarked on CASF-2016, Astex and PoseBusters.