Skip to main content
Back to projects
Bioinformatics & Genomic Analysis2025

Transcriptomic Profiling of Type 2 Diabetes — Riyadh Cohort

Whole-blood RNA sequencing of 120 participants (60 T2DM / 60 controls) from a Riyadh tertiary hospital. Identified 847 differentially expressed genes, enriched inflammatory pathways, and three hub-gene biomarker candidates for early T2DM detection in an Arabian-population cohort.

Transcriptomic Profiling of Type 2 Diabetes — Riyadh Cohort

Project highlights

120
Participants
847
DE genes
RNA-seq
Platform
0.91
Diagnostic AUC

Methods & Tools

RNA-seq (STAR / Salmon)DESeq2 / edgeRlimma-voomGSEA / fgseaPathway enrichment (clusterProfiler)R / BioconductorPython (scikit-learn)Cytoscape networks
The challenge

Discovering a reproducible transcriptomic signature for type 2 diabetes required rigorous control of false discovery across thousands of genes and a diagnostic model that generalized beyond the discovery cohort.

What we did

We built an end-to-end RNA-seq pipeline from quality control to pathway interpretation, then trained and cross-validated a parsimonious diagnostic classifier.

01

Processed RNA-seq reads (QC, STAR/Salmon alignment and quantification) for 120 participants and normalized counts with robust library-size and dispersion estimation.

02

Identified 847 differentially expressed genes with DESeq2/edgeR under Benjamini–Hochberg FDR control, and interpreted them via GSEA and clusterProfiler pathway enrichment.

03

Trained a regularized diagnostic classifier with nested cross-validation (AUC 0.91) and visualized co-expression networks in Cytoscape.

What we delivered

Reproducible RNA-seq analysis pipeline with version-controlled code.

Differential-expression and pathway-enrichment report with publication-quality figures.

Cross-validated diagnostic signature with performance and calibration metrics.

Methodology & Quality Assurance

False discovery controlled with Benjamini–Hochberg across all genes tested.

Diagnostic classifier evaluated with nested cross-validation to avoid optimism bias.

Batch effects diagnosed and adjusted before differential-expression testing.

End-to-end pipeline version-controlled and fully reproducible.

Project gallery
Transcriptomic Profiling of Type 2 Diabetes — Riyadh Cohort
Transcriptomic Profiling of Type 2 Diabetes — Riyadh Cohort
Transcriptomic Profiling of Type 2 Diabetes — Riyadh Cohort

The bioinformatics part was what worried me most — the sequencing data was big and not clean. They processed it, ran the full pipeline, and explained the results in a way I could actually put into my paper. And there were no long revisions after, which I appreciated.

O
Dr. Omar Farooq
Genomics Researcher — Riyadh Cohort

Have a project to register?

Submit your project in the dashboard and we’ll map the analysis plan, deliverables, and timeline.