Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
ML / AI methods
ML models predict antibody-antigen binding, developability, and epitope-targeted design
Learned models now span antibody affinity, developability, and TCR specificity, increasingly paired with generative design and experimental testing.
Single-cell
Analytic choices and input quality shape single-cell and spatial transcriptomics outputs
Failure modes in single-cell and spatial transcriptomics pipelines are now quantified alongside corrective and reproducibility-oriented methods.
Genomics
New tools for reference choice, variant benchmarking, structural variant calling, GWAS
Recent releases pair GPU or FPGA acceleration with benchmarking and free or open-source availability across reference, variant, and association analysis.
Single-cell
Benchmarks and models for predicting transcriptional responses to perturbations
Systematic benchmarks now test single-cell foundation models and latent representations on zero-shot and reconstruction tasks.
ML / AI methods
Learned models for tumor fraction, clonality, and foundation-model benchmarking in omics
Foundation models and Bayesian deep learning are being benchmarked against simple baselines and applied to cfDNA, flow cytometry, and single-cell data.