Top 40
View all ›PARTIAL Some sources or summaries were unavailable for this run.
BACKFILLED Code and industry signals are partial for backfilled days.
Structural bio & protein design
Protein deep learning models probed for extrapolation and physical priors
Recent preprints and publications test whether protein deep learning generalizes when physical priors, structure, or disorder context are added.
Tools & infrastructure
Configurable and federated frameworks standardize genomic variant workflows
Rare-disease genomics is moving toward standardized, privacy-compliant tooling for variant calling, sharing, and agent-driven patient data sync.
Single-cell
Long-read isoform annotations are being tested against alternative reference resources
Isoform-resolved analyses now test eQTL, TWAS, and annotation novelty against alternative transcript references rather than a fixed annotation.
ML / AI methods
Pretrained embeddings and contrastive objectives advance biomolecular binding prediction
Recent work applies transformers, contrastive learning and distillation to binding and specificity prediction from sequence and assay data.
Single-cell
Spatial geometry enters the inference step in spatial omics models
Slices with varying geometry now receive explicit geometric treatment in model inference, as in STORM, 3d-OT, and SMART.