# Identification of Potential Functional Modules and Diagnostic Genes for Crohn’s Disease Based on Weighted Gene Co-expression Network Analysis and LASSO Algorithm

## TL;DR

This study identifies key genes and modules for diagnosing Crohn’s disease using network analysis and a machine learning algorithm, offering a promising diagnostic tool.

## Contribution

A novel diagnostic model for Crohn’s disease using WGCNA and LASSO, with high accuracy and identified hub genes.

## Key findings

- 651 differentially expressed genes were linked to inflammation and leukocyte chemotaxis pathways.
- The turquoise module was associated with macrophages M2 in CD patients.
- A diagnostic model with eight hub genes achieved 0.94 AUC in training and 0.941 in validation sets.

## Abstract

Accurate diagnosis of Crohn’s disease (CD) is paramount due to its resemblance to other inflammatory bowel diseases. Early and precise diagnosis plays a pivotal role in tailoring personalized treatments, thereby enhancing the quality of life for CD patients.

Differential gene expression analysis was conducted to identify genes from the mRNA expression profiles of CD samples, followed by pathway enrichment analysis. The immune cell infiltration levels of each CD patient sample were assessed. Using weighted gene co-expression network analysis, key gene modules linked to CD were found. Hub gene identification was made easier by the construction of protein–protein interaction networks. Next, utilizing the Least Absolute Shrinkage and Selection Operator on the hub genes in the training set, a diagnostic model was created. The accuracy of the model was then confirmed using a different validation set.

Our analysis revealed 651 differentially expressed genes, enriched in leukocyte chemotaxis and inflammation-related pathways. Immunization results showed a higher abundance of T cells CD4 memory resting, macrophages M2, and plasma cells in CD patients. Weighted gene co-expression network analysis linked the turquoise module with macrophages M2. Eight hub genes (APOA1, APOA4, CYP2C8, CYP2C9, CYP2J2, EPHX2, HSD3B1, and LPL) formed the diagnostic model, exhibiting excellent diagnostic performance with area under curve values of 0.94 (training set) and 0.941 (validation set).

The CD diagnostic model, based on hub genes, shows exceptional diagnostic accuracy, providing a valuable reference for CD diagnosis.