Ferroptosis Hub Genes & Apoptosis in Ovarian Cancer

MSc Bioinformatics Dissertation | Pondicherry University | May 2024

Author: Shruti Banerjee (Reg. No. 22378042)
Supervisor: Dr. Basant K. Tiwary, Professor, Department of Bioinformatics
Degree: Master of Science in Bioinformatics

---

Overview

Ovarian cancer has one of the highest mortality rates among gynaecological malignancies, largely due to late-stage diagnosis and resistance to standard chemotherapy. This study investigates the molecular crosstalk between two distinct cell death pathways — ferroptosis and apoptosis — to identify potential therapeutic targets in ovarian cancer.

Key finding: From 7,862 differentially expressed genes, two critical ferroptosis hub genes — CDKN1A and GDF15 — were identified as co-expressed with apoptosis genes and associated with multiple carcinogenic pathways including platinum drug resistance.

---

Repository Structure

ferroptosis-ovarian-cancer-analysis/
├── README.md
├── METHODS.md
├── requirements.txt
├── thesis_shruti_banerjee.pdf
├── thesis_visualizations_fixed.py
├── thesis_visualizations.R
├── figure/
│   ├── chart1_qc_pipeline.png
│   ├── chart2_alignment_scores.png
│   ├── chart3_hub_gene_degrees.png
│   ├── chart4_kegg_pathways.png
│   ├── chart5_deg_funnel.png
│   ├── R_chart1_lollipop_hub_ranks.png
│   ├── R_chart2_coexpression_balance.png
│   └── R_chart3_gene_crosstalk.png
├── notebooks/
│   └── ferroptosis_analysis.ipynb
└── data/
    └── sample_ids.csv

---

Pipeline Overview

Raw RNA-Seq Data (ENA SRA)
        ↓
Quality Control (FastQC + Trimmomatic)
        ↓
Alignment to Reference Genome (HISAT2)
        ↓
Transcriptome Assembly (StringTie)
        ↓
Differential Gene Expression (DESeq2 in R)
        ↓
PPI Network Construction (STRING + Cytoscape)
        ↓
Gene Co-expression Network (PSYCH package in R)
        ↓
Functional Enrichment Analysis (DAVID / KEGG)
        ↓
Hub Gene Identification → CDKN1A & GDF15

---

Dataset

Parameter	Diseased Samples	Control Samples
Bioproject	PRJNA1005317	PRJNA578440
Organism	Homo sapiens	Homo sapiens
Assay type	RNA-Seq	RNA-Seq
Platform	Illumina HiSeq 300	Illumina HiSeq 300
Samples after QC	5	6
Alignment score	95–98%	96–97%

---

Tools & Technologies

Bioinformatics Pipeline

Statistical Analysis & Visualisation

Network Analysis

Databases

---

Results

1. Differential Gene Expression

Filter	Gene Count
Total DEGs identified	7,862
Significant (p < 0.05)	8,002
Highly significant (p < 0.01)	6,577
Ferroptosis + Apoptosis genes filtered	289

2. PPI Network Statistics

Network	Nodes	Interactions
All DEGs	236	886
Ferroptosis genes	174	1,469
Co-expressed hub genes	12	25

3. Ferroptosis Hub Gene Rankings

Rank	Gene	Functional Category
1	TP53	Tumour suppressor
2	HIF1A	Hypoxia response
3	EGFR	Growth signalling
4	IL6	Inflammation
4	STAT3	Inflammation
6	PARP1	DNA repair
7	GPX4	Ferroptosis regulator
8	SIRT1	Metabolism
8	MTOR	Metabolism
8	NFE2L2	Oxidative stress

4. Co-expressed Ferroptosis–Apoptosis Hub Genes

Gene	Type	Interactions	Role
CDKN1A (p21)	Ferroptosis	23	Cell cycle arrest; tumour suppressor; p53 mediator
GDF15	Ferroptosis	10	Stress-induced cytokine; cancer progression
CISD2	Ferroptosis	—	Iron-sulphur cluster; mitochondrial function
NUPR1	Ferroptosis	—	Stress response; ferroptosis resistance

5. KEGG Pathway Enrichment

Pathway	Key Genes
Pathways in cancer	CDKN1A, JUN, CYCS, CKS1B
Colorectal cancer	CDKN1A, JUN, CYCS
Small cell lung cancer	CDKN1A, CYCS, CKS1B
Renal cell carcinoma	CDKN1A, JUN
Platinum drug resistance	CDKN1A, CYCS
p53 signalling pathway	CDKN1A, CYCS
HTLV-1 infection	FOSL1, CDKN1A, JUN

---

Data Visualisations

All charts are fully reproducible. Source code in both Python and R.

Python Visualisations

Generated using Matplotlib + NumPy — see thesis_visualizations_fixed.py

Sample count through QC pipeline

HISAT2 alignment scores — diseased vs control

Hub gene interaction degrees

KEGG pathway gene involvement

DEG filtering funnel

---

R Visualisations

Generated using ggplot2 + ggalluvial — see thesis_visualizations.R

Ferroptosis hub gene rankings — lollipop chart

Publication-quality lollipop chart colour-coded by functional category.

Co-expression network composition — ferroptosis vs apoptosis

Shows that ferroptosis genes are a small but critically co-expressed minority.

Ferroptosis–apoptosis gene crosstalk — alluvial diagram

Sankey-style diagram showing CDKN1A and GDF15 bridging ferroptosis and apoptosis pathways.

---

How to Reproduce

Python:

pip install -r requirements.txt
python thesis_visualizations_fixed.py

R:

install.packages(c("ggplot2", "ggalluvial", "dplyr"))
Rscript thesis_visualizations.R

In Google Colab (R):

!apt-get install -y r-base libcurl4-openssl-dev libssl-dev libxml2-dev
!R -e "install.packages(c('ggplot2','ggalluvial','dplyr'), repos='http://cran.r-project.org')"
!Rscript thesis_visualizations.R

---

Key Biological Insights

• Platinum drug resistance: CDKN1A's involvement suggests ferroptosis induction could resensitise cisplatin-resistant ovarian tumours — a major clinical challenge.
• GPX4 (rank 7) is the primary ferroptosis gatekeeper and a promising therapeutic target in ovarian cancer.
• The p53–CDKN1A axis is a shared regulatory node between ferroptosis and apoptosis — making it a dual-pathway therapeutic target.
• The alluvial diagram reveals CDKN1A has broader apoptosis connectivity than GDF15, suggesting it plays a more central bridging role.

---

Conclusion

This study identified CDKN1A and GDF15 as key ferroptosis hub genes with co-regulatory roles in apoptosis in ovarian cancer. Their involvement in platinum drug resistance and p53 signalling opens avenues for combination therapies targeting both cell death mechanisms simultaneously.

---

Citation

Banerjee, S. (2024). Identification of ferroptosis-related hub genes and their potential relation with apoptosis in Ovarian Cancer. MSc Dissertation, Department of Bioinformatics, Pondicherry University.

---

Shruti Banerjee · banerjee.shruti1306@gmail.com · GitHub