Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
ML / AI methods
Biological priors make multi-omics deep learning interpretable
Foundation models and biology-constrained architectures are now being applied to spatial, single-cell and bulk multi-omics integration.
Systems biology
Mechanistic models link single-cell omics to biophysical and metabolic dynamics
Mechanistic and generative models are increasingly fit to single-cell, proteomic and longitudinal data to infer dynamics rather than only describe states.
Transcriptomics
Encoding spatial and relational structure in high-dimensional omics analysis
Recent spatial omics and single-cell Hi-C methods encode tissue or contact structure explicitly, rather than treating cells as independent feature vectors.
ML / AI methods
Benchmark design and confounding decide whether biological prediction models hold up
Perturbation and virtual-cell models are now judged against baselines and scientific value, not predictive accuracy alone.
ML / AI methods
Cross-subject neural biomarkers hinge on evaluation design and generalizable features
Recent preprints test whether EEG and fMRI biomarkers survive cross-subject and cross-cohort shifts, with leakage audits exposing near-chance decoding.