T2D Research

Tkachenko et al. 2024 Deciphering the Transcriptomic Landscape of T2D

Properties2
tagsreference, type-2-diabetes, pbmc, transcriptomics, single-cell
aliasesTkachenko et al. 2024 Deciphering the Transcriptomic Landscape of T2D

4 min read

Tkachenko et al. 2024 — Deciphering the Transcriptomic Landscape of Type 2 Diabetes

Preprint (medRxiv). Tkachenko A.A., Tonyan Z.N., Nasykhova Y.A., et al. (2024). Deciphering the Transcriptomic Landscape of Type 2 Diabetes: Insights from Bulk RNA Sequencing and Single-Cell Analysis. medRxiv, 2024.11.06.24316740. DOI: 10.1101/2024.11.06.24316740. Not peer-reviewed. CC-BY 4.0.

Study Design

  • Cohort: 9 T2D patients + 9 healthy controls from St. Petersburg, Russia. All participants enrolled at D.O. Ott Research Institute of Obstetrics, Gynecology and Reproductology.
  • Bulk RNA-seq: Whole blood from all 18 participants, TruSeq Stranded Total RNA library, 75 bp PE on Illumina HiSeq 2500.
  • scRNA-seq: PBMCs from a subset — 2 T2D and 2 healthy controls. All female, no family history of T2D, average age 36.
  • scRNA-seq platform: 10x Genomics Chromium, CellRanger v8.0.0, Seurat v5.1.0, Harmony integration, Louvain clustering (0.5 resolution), Azimuth annotation (pbmcref), DCATS for differential composition.
  • Bulk analysis: kallisto pseudoalignment (GENCODE v46), DESeq2 DE, clusterProfiler GO enrichment.
  • Data accessions: GSE280401 (scRNA-seq), GSE280402 (bulk RNA-seq). Subject enumeration is continuous — subjects 2 and 4 are included in both experiments.

Key Findings

scRNA-seq (PBMC)

  • 15 clusters identified by Louvain clustering: CD4+ TCM, CD4+ TCM/naive, CD8+ TEM, NK cells, CD14+ monocytes (two clusters), CD8+ naive T, intermediate B cells, naive B cells, CD16+ monocytes, MAIT, platelets, Tregs, pDCs.
  • Differential composition (DCATS): NK cells significantly less prevalent in T2D samples; CD4+ TCM and naive cells significantly more abundant in T2D. ^[extracted] (from Tkachenko et al. 2024)
  • Contrast with Gu et al. 2024: Tkachenko et al. note a different NK cell composition trend — NK depletion in T2D versus what Gu et al. observed. The authors suggest this may reflect smaller sample sizes (n=2 per group for scRNA-seq). ^[extracted] (from Tkachenko et al. 2024)

Bulk RNA-seq (Whole Blood)

  • 146 DEGs at adjusted p-value < 0.05 (71 up, 75 down) out of 74,163 genes with non-zero counts.
  • PCA with top 500 most variable genes shows samples mostly mix and do not group by disease status — majority of variance not explained by T2D. ^[extracted] (from Tkachenko et al. 2024)
  • Top 20 enriched GO terms: RNA splicing, proteasomal protein catabolic process, chromatin binding, T cell activation, mRNA processing, ribonucleoprotein complex biogenesis, regulation of proteolysis, endosome membrane, apoptotic signaling pathway, activation of immune response. ^[extracted] (from Tkachenko et al. 2024)

Clinical Characteristics

  • T2D group: mean age 66.3, 5F/4M, BMI 32, FBG 8.8 mmol/L
  • Control group: mean age 56.6, 5F/4M, BMI 24.5, FBG 4.7 mmol/L
  • Significant differences: FBG (p=0.0096), BMI (p=0.0007), WHR (p=0.01)

Related work already ingested in this wiki provides context for this preprint:

  • Tkachenko et al. 2025 — A cross-study blood RNA-seq meta-analysis (including the GSE280402 data from this 2024 paper) that found low individual-study concordance and 2065 meta-analysis DEGs enriched for neutrophil effector biology, ERAD, mTOR, oxidative stress, and RNA splicing.
  • Markelova et al. 2025 — T2D genetic heterogeneity analysis using partitioned polygenic scores across Chechen, Tatar, and Yakut populations. This is relevant same-project ancestry context, not work attributed to the Tkachenko group.

Limitations

  • Small scRNA-seq sample size (n=2 per group) limits statistical power for composition and differential expression analyses.
  • scRNA-seq subset is all female; main cohort is mixed-sex.
  • MedRxiv preprint — not yet peer-reviewed.
  • PCA shows bulk transcriptome variance is dominated by non-disease factors, suggesting batch/subject characteristics may confound T2D signals. ^[extracted] (from Tkachenko et al. 2024)
  • No ancestry-stratified analysis — the Russian cohort is treated as a single group.
  • Medication status and T2D duration not reported in detail.

Relevance to Project

  • Provides a primary-data counterpart to the meta-analysis approach of Tkachenko et al. 2025 — offers raw sequencing data (GSE280401, GSE280402) from the same research group.
  • NK cell depletion in T2D PBMCs contrasts with Gu et al. 2024, illustrating cross-study heterogeneity even at the PBMC single-cell level.
  • Bulk blood PCA (no T2D group separation by top 500 variable genes) reinforces caution about blood transcriptome signal-to-noise for the paper’s claims.

Graph View

Backlinks