Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Structural bio & protein design
Deep learning biomolecular models coupled to experimental and geometric constraints
Recent work injects XL-MS, HDX-MS, surface geometry, and oracle-budgeted search into pretrained protein and ligand design models.
Single-cell
Computational frameworks link perturbations to single-cell and spatial readouts
Perturbation screens and spatial atlases now generate multimodal, terabyte-scale readouts that require scalable, integrated computational analysis pipelines.
ML / AI methods
Deep sequence models predict gene expression, splicing and variant effects
Sequence-to-function models are now being extended to personal genomes, tumor promoter variants, and multi-tissue TWAS fine-mapping.
ML / AI methods
Learned representations recover structure in omics and pathology data across platforms
Benchmarks and graph, contrastive and flow-based models now test transferability across platforms, cohorts and modalities in omics prediction.
Transcriptomics
Spatial omics methods map niche structure, communication and perturbation responses
Recent items pair spatial CRISPR screens and niche-level null models with graph-based causal inference, extending single-cell data into tissue context.