Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Structural bio & protein design
Geometry-aware models refine antibody and peptide binding and stability prediction
Recent preprints and publications apply residual correction, flow matching and graph attention to structure-derived binding, stability and epitope tasks.
Single-cell
Single-cell methods model perturbation responses, spatial maps, and technical dropout
Current single-cell work pairs perturbation modeling and spatial integration with explicit correction of technical noise and dropout.
ML / AI methods
Transferable deep learning models extend omics integration, prognosis and deconvolution
Foundation and transformer models are now being transferred across proteomics, cytometry, cfDNA and single-cell datasets.
ML / AI methods
Transformer and foundation models map single-cell states and perturbation responses
State and DECANT predict single-cell perturbation effects, while scProtoTransformer and ELISA extend reference mapping and expression-grounded querying.
Single-cell
Deep models fuse histology, spatial and single-cell omics into structured embeddings
Heterogeneous single-cell, spatial and histology data now support interpretable, generative models that encode gene-network and niche structure explicitly.