Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
ML / AI methods
Leakage-free validation and benchmarking reshape omics ML biomarker discovery
Omics-based predictive and biomarker models increasingly require leakage-free validation and explicit benchmarks against established methods.
Single-cell
Calibrating single-cell and spatial inference after imputation and spatial dependence
Imputed spatial transcriptomes, multi-omics similarity, and single-cell normalization now demand explicit error modeling for valid inference.
Genomics
Multiancestry methods extend variant association mapping to cell types and deep models
Ancestry-diverse and cell-type-resolved datasets now test whether eQTL, fine-mapping, and sequence models generalize beyond European bulk-tissue settings.
ML / AI methods
Protein language models predict immune recognition across TCR, pMHC and B-cell epitopes
Recent work pairs protein language models with experimental readouts, from pMHC stability to antigen-specific T cell calling.
Genomics
Reproducible pipelines and benchmarks for long- and short-read sequencing analysis
New open tools and donor-resolved benchmarks target reproducibility and error profiling across long- and short-read sequencing workflows.