Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Single-cell
Graph and pathway priors guide multimodal representations of single-cell and patient data
Recent preprints couple regulatory networks, pathway graphs and multi-omics inputs into single embeddings for predicting response and disease states.
ML / AI methods
Protein language models extend to interactions, modifications, and modified peptides
Pretrained sequence representations are now applied to protein-protein interactions, PTM syntax, enzymes, and peptide developability in recent work.
ML / AI methods
Probing and auditing what biological sequence foundation models actually encode
Pre-registered and contamination-aware evaluations now test whether DNA, RNA, and single-cell foundation models outperform strong controls.
Genomics
Chromatin and computational methods resolve genomic and RNA fragmentation patterns
Long-read, single-cell and cfDNA datasets now expose fragmentation, copy number and epigenomic states at high resolution across cancer and development.
Single-cell
Interpretable pathway, network and cell-cell features from high-dimensional omics data
Recent preprints and one published study add reliability testing, benchmarking and multiscale spatial scoring to omics network and single-cell methods.