Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Single-cell
Histology-conditioned models predict spatial gene expression from H&E morphology
Recent preprints and publications recast H&E-to-transcriptome prediction as generative modeling over transcriptional programs and enhance sparse spatial data…
ML / AI methods
Protein and RNA language models predict binding sites, PTMs and mutation energetics
Pretrained ESM2, RNA-FM, Evo, and BiRNA-BERT embeddings are now being fused with specialized architectures for residue-level prediction tasks.
ML / AI methods
Histology representations are being aligned with spatial gene expression and language
Recent preprints pair H&E histology with spatial transcriptomics for gene expression prediction, cross-modal alignment, and prognostic analysis.
Tools & infrastructure
New tools link RNA isoforms and somatic variants to proteomic and genomic readouts
Recent tools and benchmarks address isoform, fusion, and rare-variant inference from long-read, mass-spec, and sequencing data.
Single-cell
Generative and linear models predict single-cell perturbation transcriptional responses
Generative perturbation models and single-cell foundation models are now evaluated on held-out drugs, cell types and token budgets.