{
  "id": 9851413,
  "title": "fastACCORD enables ultrahigh-dimensional partial correlation modeling for multi-omic data integration",
  "url": "https://urgent.news/2026/09/25/fastaccord-enables-ultrahigh-dimensional-partial-correlation-modeling",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-25T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.20.753046v1?rss=1"
  },
  "original_language": "en",
  "account": "Gene co-expression networks result from transcription factor and epigenetic influences on gene regulation. Conventional methods for statistical inference on hundreds of thousands of molecular features are computationally overwhelming. The fastACCORD framework addresses this challenge with a scalable, CPU and CUDA-enabled GPU-compatible implementation. fastACCORD uses row-separable optimization, \\\\ell2 stabilization, and a semismooth Newton solver for efficient partial correlation modeling. Applied to matched transcriptomic and methylomic profiles from 16 TCGA cancer types, fastACCORD revealed joint networks containing 300,000 molecular features per cancer. These networks showed cancer-specific methylation-expression dependencies, including repression of metabolic genes. Methylation-adjusted co-expression networks were sparser than those estimated from mRNA data alone but enriched for ChIP-seq-supported TF-target relationships and curated co-regulon annotations. TF-target subnetworks identified cancer-specific architectures related to lineage identity, oncofetal reactivation, and tumor microenvironment-associated programs. Overall, fastACCORD proves to be a practical framework for ultrahigh-dimensional multi-omic partial correlation modeling, demonstrating the benefits of joint modeling of transcriptomic and epigenomic measurements in interpreting gene co-expression networks.",
  "summary": "Gene co-expression networks reflect transcription factor-associated regulation together with epigenetic influences such as DNA methylation, chromatin states, and histone modifications. To distinguish gene-gene dependencies that persist after accounting for methylation covariation, large-scale statistical inference conditioning on hundreds of thousands of molecular features is necessary, but the…",
  "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."
}