Skip to content

daRk8238/Pan_diabetes_complication_analysis

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
scripts
scripts
 
 
README.md
README.md
 
 
environment.yml
environment.yml
 
 
requirements.txt
requirements.txt
 
 

Repository files navigation

Transcriptomic Profiling and Regulatory Network Analysis of Ten Metabolic Transporters Across Five Diabetic Complications

A Multi-Dataset, Twelve-Phase GEO Bioinformatics Study

Python License: MIT GEO Status

Overview

This repository contains the complete analysis pipeline for a twelve-phase integrative bioinformatics investigation of ten key metabolic transporters across five major diabetic complications:

Complication Abbreviation Tissue
Diabetic cardiomyopathy DCM Cardiac tissue / myocardial biopsy
Diabetic nephropathy DN Glomerular / renal cortex tissue
Diabetic retinopathy DR Retinal tissue
Diabetic peripheral neuropathy DPN Sural nerve
Diabetic atherosclerosis / vasculopathy DAD Aortic / vascular smooth muscle tissue

The ten target transporters: SLC2A4 (GLUT4) · SLC5A2 (SGLT2) · CD36 · AGER (RAGE) · SLC16A3 (MCT4) · TLR4 · ATP2A2 (SERCA2) · SLC8A1 (NCX1) · SLC2A1 (GLUT1) · SLC27A1 (FATP1)

Study Design

11 GEO Datasets (118 disease · 76 control · 194 total samples)
                        │
        ┌───────────────▼───────────────┐
        │      TRANSCRIPT LEVEL         │
        │  Phase 1: Data Acquisition    │
        │  Phase 2: Diff. Expression    │
        │  Phase 3: Pan-Complication    │
        └───────────────┬───────────────┘
                        │
        ┌───────────────▼───────────────┐
        │       NETWORK LEVEL           │
        │  Phase 4: WGCNA               │
        │  Phase 5: GO/KEGG/GSEA        │
        │  Phase 6: PPI Networks        │
        └───────────────┬───────────────┘
                        │
        ┌───────────────▼───────────────┐
        │      REGULATORY LEVEL         │
        │  Phase 7: ceRNA Network       │
        │  Phase 8: TF Activity         │
        └───────────────┬───────────────┘
                        │
        ┌───────────────▼───────────────┐
        │       CELLULAR LEVEL          │
        │  Phase 9: Immune Deconv.      │
        └───────────────┬───────────────┘
                        │
        ┌───────────────▼───────────────┐
        │    TRANSLATIONAL LEVEL        │
        │  Phase 10: Diagnostics        │
        │  Phase 11: Drug Repurposing   │
        │  Phase 12: Mendelian Rand.    │
        └───────────────┬───────────────┘
                        │
        Pan-Complication Mechanisms · Diagnostic Biomarkers
        Drug Repurposing Candidates · Causal Genomic Evidence

Repository Structure

transporter-diabetic-complications/
│
├── README.md                          # This file
├── environment.yml                    # Conda environment specification
├── requirements.txt                   # pip package list
│
├── scripts/
│   ├── 01_data_acquisition.py         # Phase 1: GEO download & QC
│   ├── 02_differential_expression.py  # Phase 2: DEA (Welch t-test, BH-FDR)
│   ├── 03_pan_complication.py         # Phase 3: Pearson correlation overlap
│   ├── 04_wgcna.py                    # Phase 4: Co-expression network (kME)
│   ├── 05_enrichment_gsea.py          # Phase 5: GO/KEGG/GSEA enrichment
│   ├── 06_ppi_networks.py             # Phase 6: STRING PPI network analysis
│   ├── 07_cerna_network.py            # Phase 7: ceRNA regulatory network
│   ├── 08_tf_activity.py              # Phase 8: VIPER-style TF inference
│   ├── 09_immune_infiltration.py      # Phase 9: ssGSEA immune deconvolution
│   ├── 10_diagnostics.py              # Phase 10: LASSO + Random Forest
│   ├── 11_drug_repurposing.py         # Phase 11: CMap-style connectivity
│   └── 12_mendelian_randomisation.py  # Phase 12: Two-sample MR (IVW/Egger/WM/WMo)
│
├── outputs/
│   ├── 02_DEA/                        # Volcano plots, fold-change tables
│   ├── 03_PanComplication/            # Correlation matrices, UpSet plots
│   ├── 04_WGCNA/                      # Module membership, eigengene plots
│   ├── 05_Enrichment/                 # KEGG dotplots, GSEA bar charts
│   ├── 06_PPI/                        # Network graphs, hub gene tables
│   ├── 07_ceRNA/                      # Tripartite network, centrality stats
│   ├── 08_TF/                         # Activity heatmaps, regulon scores
│   ├── 09_Immune/                     # ssGSEA delta tables, heatmaps
│   ├── 10_Diagnostics/                # ROC curves, feature importance
│   ├── 11_Drugs/                      # Connectivity scores, drug tables
│   └── 12_MR/                         # Forest plots, causal estimates
│
├── figures/
│   ├── Figure1_Study_Pipeline.png
│   ├── Figure2_PanComplication_Heatmap.png
│   ├── Figure3_ceRNA_Network.png
│   ├── Figure4_Immune_Deconvolution.png
│   └── Figure5_MR_Forest_Plots.png
│
└── data/
    └── README_data.md                 # Data download instructions (see below)

Datasets

All data are publicly available from NCBI GEO. No data files are stored in this repository. Download instructions are in data/README_data.md.

GEO Accession Complication Organism Tissue N Disease N Control
GSE123975 DCM Mouse Cardiac tissue 6 6
GSE21610 DCM Human Myocardial biopsy 60 8
GSE30528 DN Human Glomerular tissue 9 13
GSE1009 DN Human Renal cortex 3 3
GSE111154 DN Human Kidney cortex 4 4
GSE104948 DN Human Glomerular tissue 12 21
GSE60436 DR Human Retinal tissue 6 3
GSE95849 DPN Human Sural nerve 6 12
GSE121487 DAD Mouse Aortic tissue 3 3
GSE57329 DAD Mouse Aortic tissue 3 3
GSE66280 DAD Human Vascular smooth muscle 6 6

Total: 118 disease samples · 76 control samples · 194 total

Installation and Setup

Prerequisites

1. Clone the repository

git clone https://github.com/[username]/transporter-diabetic-complications.git
cd transporter-diabetic-complications

2. Create and activate the conda environment

conda env create -f environment.yml
conda activate hmox

3. Verify installation

python -c "import GEOparse, gseapy, sklearn, networkx; print('All packages loaded successfully')"

Computational Environment

System specifications (original analysis)

Parameter Value
Operating system Ubuntu 22.04 LTS
Machine Dell workstation
Python version 3.10.x
Conda environment hmox
Random seed (all stochastic analyses) 42

Core Python packages

Package Version Purpose
GEOparse 2.0.4 GEO dataset acquisition and SOFT file parsing
gseapy 1.0.4 GO/KEGG enrichment and GSEA pre-ranked analysis
scikit-learn 1.3.0 LASSO logistic regression, Random Forest, cross-validation
networkx 3.1 PPI network construction and centrality analysis
numpy 1.24.3 Numerical computation
pandas 2.0.2 Data manipulation and tabular analysis
scipy 1.11.1 Statistical tests (Welch t-test, Mann-Whitney U)
statsmodels 0.14.0 Multiple testing correction (Benjamini-Hochberg FDR)
matplotlib 3.7.1 Figure generation
seaborn 0.12.2 Statistical visualisation
requests 2.31.0 STRING API and MyGene.info REST API queries
biopython 1.81 Biological data handling

External databases accessed (no local installation required)

Database Version / Date Purpose
NCBI GEO Accessed 2024 Microarray dataset source
STRING v12.0 Protein–protein interaction networks (score ≥ 400)
MyGene.info REST API Probe-to-gene symbol secondary mapping
miRTarBase v9.0 miRNA–target strong-evidence interactions
ENCORI Current lncRNA–miRNA sponge interactions (≥3 CLIP experiments)
lncBase v3 lncRNA–miRNA validated interactions
TRRUST v2 Transcription factor regulon (confidence A/B)
DoRothEA v1 Transcription factor regulon (confidence A/B)
DrugBank v5.1 Drug target database for repurposing analysis
GTEx v8 cis-eQTL instruments for Mendelian randomisation
DIAGRAM 2022 T2DM GWAS summary statistics
CKDGen 2021 eGFR/DN GWAS summary statistics

Reproducing the Analysis

Run phases sequentially. Each script reads inputs from the previous phase's outputs directory.

# Phase 1 — Data acquisition and quality control
python scripts/01_data_acquisition.py

# Phase 2 — Differential expression analysis
python scripts/02_differential_expression.py

# Phase 3 — Pan-complication overlap
python scripts/03_pan_complication.py

# Phase 4 — WGCNA co-expression network
python scripts/04_wgcna.py

# Phase 5 — GO/KEGG enrichment and GSEA
python scripts/05_enrichment_gsea.py

# Phase 6 — PPI network construction and analysis
python scripts/06_ppi_networks.py

# Phase 7 — ceRNA regulatory network mapping
python scripts/07_cerna_network.py

# Phase 8 — Transcription factor activity inference
python scripts/08_tf_activity.py

# Phase 9 — Immune cell infiltration (ssGSEA)
python scripts/09_immune_infiltration.py

# Phase 10 — Diagnostic biomarker modelling
python scripts/10_diagnostics.py

# Phase 11 — Drug repurposing (CMap-style)
python scripts/11_drug_repurposing.py

# Phase 12 — Mendelian randomisation
python scripts/12_mendelian_randomisation.py

Note: Phases 1–3 require an active internet connection for GEO downloads and MyGene.info API queries. Phases 4–12 run entirely on local data. All stochastic analyses (WGCNA, Random Forest, permutation testing) use random_seed = 42 to ensure full reproducibility.

Key Findings Summary

Finding Value
Pan-complication transcriptional driver CD36 (significant in DN, DR, DPN)
Highest fold-change TLR4 (log2FC = +3.88, DPN)
Strongest WGCNA hub TLR4 (kME = 0.976, DPN module)
Most consistently suppressed SERCA2 (all 5 complications)
Dominant ceRNA bridge miR-21-5p (betweenness = 0.428)
Pan-complication lncRNA MALAT1 (all 5 complications)
Perfect DPN diagnostic GLUT4 (AUC = 1.000)
Perfect DAD diagnostic SGLT2 (AUC = 1.000)
Pan-complication drug Epalrestat (4/5 complications)
Causal DN risk factor (eQTL-MR) TLR4 (β = +0.073, p = 0.006)
Causal DN risk factor (eQTL-MR) CD36 (β = +0.070, p = 0.008)
Causal DN protector (eQTL-MR) SGLT2 (β = −0.070, p = 0.013)
Causal DCM driver (eQTL-MR) SERCA2 (β = −0.085, p = 0.005)

Citation

If you use this code or data in your research, please cite:

Adegboyega, B. B., et al. (2026). Transcriptomic Profiling and Regulatory Network
Analysis of Ten Metabolic Transporters Across Five Diabetic Complications: A
Multi-Dataset, Twelve-Phase GEO Bioinformatics Study. Diabetologia.
[DOI to be added upon publication]

Data Availability

All GEO datasets are publicly available at https://www.ncbi.nlm.nih.gov/geo/. No primary data were generated in this study. GWAS summary statistics used in Phase 12 are available from:

Licence

This project is licensed under the MIT Licence. See LICENSE for details.

You are free to use, modify, and distribute this code for academic and non-commercial purposes with appropriate attribution.

Contact

Babatunde Benjamin Adegboyega Department of Genetics and Biotechnology University of Medical Sciences, Ondo, Nigeria 📧 Adegboyegabb@gmail.com 🔗 LinkedIn

Supervisor: Dr. Ifeoluwa Oyeyemi — ioyeyemi@unimed.edu.ng

Acknowledgements

The authors thank the research groups who deposited the eleven GEO datasets used in this study, the DIAGRAM Consortium, the CKDGen Consortium, and the GTEx Consortium for making GWAS and eQTL summary statistics publicly available.

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

1 star

Watchers

1 watching

Forks

1 fork
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages