Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Transcriptomics
Graph neural networks resolve spatial niches and domains in tissue transcriptomics
Recent preprints and papers apply graph and hypergraph learning to spatial transcriptomics for domain detection, imputation, and niche discovery.
ML / AI methods
Benchmarks test whether omics machine learning models generalize beyond reported…
Reported gains on single-cell and multi-omics tasks are now being tested across datasets, pipelines, and populations to assess generalizability.
Tools & infrastructure
Error-aware variant calling for low-frequency mutations in cfDNA and workflow tooling
Background error models for low-frequency cfDNA variants appear alongside new reproducible Snakemake and Nextflow variant detection pipelines.
Single-cell
Single-cell and spatial omics methods that expose interpretable inferential structure
Recent work pairs single-cell and spatial omics inference with explicit validation, calibrated statistics, and multiple interpretable solutions.
ML / AI methods
Benchmarks reveal generalization limits of ML models across biological prediction tasks
Held-out and frozen-representation benchmarks are exposing unstable model rankings and generalization gaps in biological ML models.