Identification of Potent Biomarkers for Prostate Cancer Through AR, Mapk and M-TOR Signaling Pathways Mining

Overview

This repository contains all scripts, results, and documentation related to the bachelor's thesis project focused on identifying biomarkers for prostate cancer by analyzing the AR (Androgen Receptor), MAPK, and m-TOR signaling pathways using microarray datasets. The study integrates bioinformatics, pathway analysis, and statistical approaches to discover genes that may serve as potential therapeutic targets.

Project Components

The repository is structured as follows:

1. Scripts

• view_dataset.R – Loads and explores datasets using the GEOquery package.
• parse_CEL.R – Preprocesses .CEL files from GEO datasets and normalizes the data.
• packages.R – Installs and loads necessary Bioconductor packages.
• WBDEGS.R – Runs WB-DEGS (a Shiny app for differential gene expression analysis).
• get_supplement.R – Downloads supplementary GEO datasets.

2. Data and Results

• results_annotation.xlsx – Annotated results of significant genes identified.
• results_STRING.xlsx – STRING network analysis results for gene interactions.
• results.docx – Summary of STRING and GeneMANIA interactions, listing key biomarkers.
• results.xlsx – Comprehensive results including differentially expressed genes (DEGs) across datasets.

3. Documentation

• methods.pdf – Details the methodology, including preprocessing, statistical analysis, and pathway mapping.
• datasets.pdf – Lists all GEO datasets used for analysis, including descriptions and accession numbers.
• thesis.pdf – The full bachelor's thesis report.

Methodology

1. Data Collection

   • Microarray gene expression datasets were retrieved from GEO (NCBI Gene Expression Omnibus).
   • Selected datasets targeted AR, MAPK, and m-TOR pathways.

2. Preprocessing & Normalization

   • Background correction and quantile normalization were performed using affy and limma packages.
   • RMA normalization was applied to preprocess .CEL files.

3. Differential Expression Analysis

   • Conducted using GEO2R, MeV (MultiExperiment Viewer), and WB-DEGS.
   • Applied statistical tests: t-test, linear models, twilight, and SAM (Significance Analysis of Microarrays).
   • Identified overexpressed and underexpressed genes in prostate cancer samples.

4. Pathway & Network Analysis

   • STRING and GeneMANIA were used to analyze gene interactions.
   • Identified intra-pathway and inter-pathway interactions between AR, MAPK, and m-TOR genes.
   • Key biomarkers were selected based on network connectivity and literature evidence.

Key Findings

• 13 candidate genes identified as potential biomarkers for prostate cancer.
• Hub genes found with strong interactions across pathways:
  • AR Pathway: PRKACB, CDK1, EIF5B
  • MAPK Pathway: EDN1, RPS6, SERBP1
  • m-TOR Pathway: RPL23, RPS20, UCHL5
• Inter-pathway connections suggest interactions between AR, MAPK, and m-TOR genes in prostate cancer progression.

How to Use

1. Install Required Packages
   r source("http://bioconductor.org/biocLite.R") biocLite(c("GEOquery", "affy", "limma", "gcrma", "shiny"))
2. Run Data Preprocessing
   r source("parse_CEL.R")
3. Perform Differential Expression Analysis
   r source("WBDEGS.R")
4. Explore Pathway Interactions
   
   • Open results_STRING.xlsx and results_annotation.xlsx to review gene interactions.

Authors & Acknowledgments

• Abhinav Mishra, Nimisha Asati
• Supervisor: Dr. Tiratha Raj Singh
• Institution: Jaypee University of Information Technology, Waknaghat
• Year: 2017

License

This project is open-source under the MIT License.

Reference

A. Mishra and N. Asati, Identification of Potent Biomarkers for Prostate Cancer Through AR, MAPK, and m-TOR Signaling Pathways Mining. Solan, HP: Jaypee University of Information Technology, 2017.