Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
ML / AI methods
Structured predictive models of cells, drugs and molecules aim for interrogable outputs
Recent work pairs pretrained cell and molecule models with explicit mechanistic structure to support counterfactual and attribution-based analysis.
Single-cell
Methods refine cell-level inference from spatial and single-cell transcriptomes
Multi-sample spatial and single-cell studies now demand calibrated power, decontamination, and causal path methods with explicit validation.
Structural bio & protein design
Deep learning predicts ligand binding, selectivity, and druggable pockets across targets
Learned models coupled to docking and MD now screen compound libraries of up to 1.58M compounds for diverse targets.
ML / AI methods
Explainable models over biological priors for cancer and toxicity prediction
Explainable models now integrate epigenomic, network and cell-profiling data across gene therapy, drug repurposing and cancer subtyping.
Structural bio & protein design
Physics-informed generative models steer protein, antibody, and conformational design
Recent preprints and publications couple protein language models and AlphaFold-class predictors with physics and experimental feedback for design.