T2D Research

T2D Islet Transcriptome Biomarker Discovery

Properties2
tagsconcept, type-2-diabetes, transcriptomics, biomarker, method, pancreatic-islet
aliasesT2D Islet Transcriptome Biomarker Discovery

3 min read

T2D Islet Transcriptome Biomarker Discovery

Huang et al. 2022 demonstrated a multi-step bioinformatics pipeline for T2D biomarker discovery using pancreatic islet transcriptomics. This approach is conceptually related to but distinct from PBMC-based transcriptome profiling.

Pipeline Steps

  1. Bulk RNA-seq DEG analysis (GSE41762, islet tissue): 111 DEGs via limma.
  2. Functional enrichment (GO/KEGG): hormone secretion, JAK-STAT, Ras signaling.
  3. LASSO biomarker selection: 6 genes → 4 with AUC >0.8 (SLC2A2, SERPINF1, RASGRP1, CHL1).
  4. Nomogram construction: Combined panel AUC = 0.902.
  5. scRNA-seq cell clustering (E-MTAB-5061, pancreatic islets): 13 clusters → 6 cell types.
  6. Biomarker localization: SERPINF1 identified in fibroblasts.
  7. SCENIC TF analysis: NR2F2 identified as TF regulating SERPINF1-high fibroblasts.
  8. Target gene intersection: 18 NR2F2 target genes overlap with T2D DEGs.

Tissue vs. Blood Context

This pipeline operates on pancreatic islet tissue — the disease-affected organ — rather than peripheral blood or PBMCs. Islet transcriptomics captures tissue-level pathology (β-cell dysfunction, fibrosis), while PBMC transcriptomics captures systemic immune remodeling. The two approaches provide complementary views of T2D biology.

Tang et al. 2026: scRNA-seq + Machine Learning Pipeline

Tang et al. 2026 demonstrated an alternative pipeline using single-cell RNA-seq as the starting point rather than bulk RNA-seq:

  1. scRNA-seq clustering (GSE221156, pancreatic islet): 19 subclusters → 10 cell types via Seurat + Harmony + SingleR.
  2. DEG analysis: 455 DEGs via limma.
  3. LASSO biomarker selection: 14 genes from DEGs → 4 with non-zero coefficients: PNLIP, BUB1, CTSB, NAMPT.
  4. ROC validation (GSE29221, skeletal muscle): AUCs 0.694–0.931.
  5. Cell-cell communication: CellChat analysis identifying Alpha/Beta cells as signaling hubs (MK, SPP1 pathways).
  6. Clinical qRT-PCR validation: Peripheral blood serum from a Chinese cohort (15 T2DM, 20 controls).

Key Differences from Huang et al. 2022 Pipeline

  • Input data: Tang uses scRNA-seq directly for DEG discovery; Huang uses bulk RNA-seq for DEGs then scRNA-seq for cell-type localization.
  • Validation strategy: Tang validates across tissues (islet→muscle→blood); Huang validates within islet via nomogram.
  • Cell-level analysis: Tang uses CellChat for intercellular communication; Huang uses SCENIC for TF regulation.
  • Biomarker panel: No overlap in gene membership, suggesting pipeline-specific sensitivity.

Relationship to Other T2D Transcriptome Studies

  • Blood transcriptome meta-analysis addresses cross-cohort replication challenges not tested here.
  • Single-cell PBMC profiling captures immune cell heterogeneity, while this islet study captures tissue-resident cell heterogeneity.
  • Unlike Tkachenko et al. 2025, this study does not assess replication across multiple cohorts.
  • Tang et al. 2026 provides an scRNA-seq-first alternative to the bulk-first Huang et al. pipeline, with a different biomarker panel and cross-tissue validation.

Graph View

Backlinks