{
  "id": 11255059,
  "title": "Scalable saturation mutagenesis reveals gene regulatory architecture and rare variant effects",
  "url": "https://urgent.news/2026/10/01/scalable-saturation-mutagenesis-reveals-gene-regulatory-architecture",
  "topic": "science",
  "section": "Science",
  "published": "2026-10-01T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.30.755794v1?rss=1"
  },
  "original_language": "en",
  "account": "Long-context sequence-to-function models can predict regulatory variant effects at the nucleotide level, prompting large-scale mutational interrogation. Traditional in silico saturation mutagenesis (ISM) analyzes mutations individually, necessitating millions of model evaluations per gene and billions to trillions for entire genomes. Multi-ISM, a new framework, addresses this by reframing ISM as a sparse recovery problem. Multi-ISM produces mutational maps using 45 times fewer model evaluations than exhaustive single-variant ISM, yet matches or surpasses its accuracy on variant-effect benchmarks. The framework is architecture-agnostic and applicable to large-scale sequence-to-function models. Researchers applied Multi-ISM to 5,000 protein-coding genes, with 3,317 being disease-associated genes, obtaining base-pair-level, tissue-specific attribution maps across 500-kb windows. These maps helped prioritize enhancers, identify genes tied to specific cell types, and reveal regulatory complexity. The summary data from these maps demonstrated that genes with tighter constraints saw smaller predicted mutational impacts. By combining Multi-ISM predictions into gene-level rare-variant burdens, personalized expression prediction outperformed common-variant elastic net models, with the most significant improvements observed at expression outliers. Multi-ISM transforms nucleotide-resolution interpretation of long-context sequence models from a specialized task to a routine computation, enabling the mapping of new architectures, functional readouts, and cellular contexts as they emerge.",
  "summary": "Long-context sequence-to-function models enable nucleotide-resolution prediction of regulatory variant effects, motivating comprehensive mutational interrogation across broad genomic contexts. Yet conventional in silico saturation mutagenesis (ISM) scores mutations one at a time, requiring millions of model evaluations for a single gene and billions to trillions at genome scales. Here, we…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}