{
  "id": 9865184,
  "title": "HR-FRGS: A Novel Biomarker Discovery Protocol using Dual Layer Hypergraph Learning for NGS RNA Sequence Data",
  "url": "https://urgent.news/2026/09/25/hr-frgs-a-novel-biomarker-discovery-protocol-using-dual-layer",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-25T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.21.753079v1?rss=1"
  },
  "original_language": "en",
  "account": "The research paper, \"HR-FRGS: A Novel Biomarker Discovery Protocol using Dual Layer Hypergraph Learning for NGS RNA Sequence Data,\" introduces a new framework called Hypergraph Regularised Fuzzy Rough Gene Selection (HR-FRGS) to address challenges in discovering biomarkers from high-dimensional RNA sequencing data. Conventional methods like differential expression analysis and weighted gene co-expression network analysis often yield false-positive interactions and are vulnerable to transitivity-driven noise and arbitrary thresholds for feature selection. To overcome these limitations, the HR-FRGS pipeline combines protein clustering, entropy filtering, and PPI-pruned co-expression interactions to create a dual-layer hypergraph.\n\nA novel hypergraph-based scoring method, incorporating Random Walk with Restart diffusion and Hypergraph Betweenness Centrality, is employed to identify genes that are central in the biological network. An autonomous fuzzy rough set selection further eliminates the need for arbitrary threshold dependency in gene selection. The HR-FRGS framework was applied to four TCGA cancer cohorts, including Lung Adenocarcinoma, Head and Neck Cancer, Kidney Clear Cell Carcinoma, and Colon Cancer. This resulted in compact biomarker panels that reduced over 60,000 transcripts.\n\nThe effectiveness of the biomarker panels was validated using classical Machine Learning algorithms with cross-validation, demonstrating that they matched or surpassed the classification performance of the full transcriptome (AUC of 0.99 and MCC of 0.94). Furthermore, the biomarker panels exhibited generalizability, with a minimum Matthews correlation coefficient (MCC) of 0.893 across the cohorts. Benchmarking against baseline methods and an ablation study highlighted the contributions of key components of the algorithm. The functional enrichment analysis revealed established pan-cancer hallmarks and cohort-specific oncogenic mechanisms, confirming the biological relevance of the identified genes.",
  "summary": "Biomarker discovery from high-dimensional RNA sequencing data remains challenging. Conventional methods such as differential expression analysis, pairwise protein-protein interaction networks, and weighted gene co-expression network analysis suffer from false-positive interactions, transitivity-driven noise, and arbitrary thresholds for feature selection. Addressing these limitations, this study…",
  "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."
}