# Unraveling gene interaction networks in colorectal cancer and inflammatory bowel disease via a novel hybrid radial basis function network

## TL;DR

This study uses a new machine learning method to uncover gene interactions in colorectal cancer and inflammatory bowel disease, identifying key genes and pathways involved.

## Contribution

A novel Hybrid Radial Basis Function Network approach for gene clustering is introduced, showing improved performance over conventional methods.

## Key findings

- APC, SMAD4, and MSH2 were identified as critical genes in CRC and IBD gene networks.
- The Hybrid RBF Network outperformed traditional clustering methods with a Silhouette Score of 0.70 and other high metrics.
- NLRP3 and PYCARD interactions suggest inflammasomes may link chronic inflammation to cancer development.

## Abstract

Colorectal cancer (CRC) is a significant global health challenge, closely linked with inflammatory bowel disease (IBD). Understanding the genetic and molecular underpinnings of CRC and its association with IBD is critical for early diagnosis and personalized treatment. This study introduces a novel Hybrid Radial Basis Function (RBF) Network approach for gene clustering to uncover key genetic interactions and pathways associated with these conditions. Gene datasets related to CRC and IBD were retrieved from public databases, including OMIM and Entrez Gene. Functional and structural analyses of these genes were conducted using bioinformatics tools such as STRING and GeneMania. A Hybrid RBF Network clustering methodology was employed to analyze gene sequence similarities, leveraging density thresholds, Gaussian functions, and clustering resolution parameters for optimal performance. The clustering quality was evaluated using metrics like the Silhouette Score, Calinski–Harabasz Index, and Davies–Bouldin Index. The study identified central genes such as APC, SMAD4, and MSH2 as critical nodes in the gene interaction network, emphasizing their role in CRC and IBD pathogenesis. The clustering methodology demonstrated superior performance (Silhouette Score: 0.70; Calinski–Harabasz Index: 30.5; Davies-Bouldin Index: 0.50) compared to conventional techniques. Furthermore, interactions between NLRP3 and PYCARD highlighted the potential involvement of inflammasomes in linking chronic inflammation to carcinogenesis. The proposed Hybrid RBF Network approach provides a robust framework for gene clustering and provides new insights into the genetic basis of CRC and IBD. Our work highlights the transformative potential of machine learning and bioinformatics in advancing genomic research and precision medicine.