Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Systems biology
Graph and structure-aware models encode biological priors for omics-based prediction
Explicit structural priors, from metabolic topology to regulatory networks, are now extended into single-cell and cohort-level prediction.
Single-cell
Generative models couple single-cell and spatial readouts with external priors
Perturbation prediction, histology-to-expression mapping, and mosaic integration now use learned priors to constrain single-cell inference.
ML / AI methods
Protein language models are benchmarked and adapted for mutation effect prediction
Recent preprints stress-test protein language models for mutation, stability, and toxicity tasks, emphasizing evaluation design and efficient adaptation.
ML / AI methods
Foundation models benchmarked on viral fitness, microbial genomes and pathogenicity
Genome and protein language models are now being tested against alignment baselines and experimental fitness data across viral, microbial, and protein tasks.
Genomics
Long-read assemblies resolve chromosome architecture, repeats and sex chromosomes
Nanopore and HiFi workflows now yield gap-free, haplotype-resolved assemblies from non-model invertebrates, fish and plants.