Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Single-cell
Biological priors are being built into single-cell and spatial transcriptomics models
Recent preprints and papers embed regulatory, pathway, and language-model gene priors into deep models to anchor predictions in known biology.
ML / AI methods
ML models for omics and drug response prediction gain calibration and interpretability
Recent work tests whether simple, calibrated, or interpretable models match complex architectures for omics and drug-response prediction.
Structural bio & protein design
Structure-conditioned deep models target protein design, ranking, and kinetics
Structure-derived inputs and targets now feed generative, ranking, and kinetics models across antibody, nanobody, peptide, and protein engineering tasks.
ML / AI methods
Benchmark choices and inductive biases shape ML drug-response and design models
Held-out chemistry and out-of-distribution splits are exposing how metric choice and identifiability limit model claims in drug discovery.
Epigenomics
Learned models predict regulatory function from sequence, methylation, and chromatin
Recent preprints and publications extend sequence-to-function and chromatin state models to m6A, splicing, and enhancer identification.