# Research on the diagnosis model of osteoarthritis based on methylation-related genes using machine learning algorithms

## TL;DR

This study builds a machine learning model using methylation-related genes to diagnose osteoarthritis and evaluates its effectiveness.

## Contribution

A novel diagnostic model for osteoarthritis using methylation-related genes and machine learning is developed and validated.

## Key findings

- 11 differentially methylated genes were identified as key features for the diagnostic model.
- The model achieved high accuracy with AUC scores of 0.96 and 0.93 in two datasets.
- Gene enrichment analysis revealed biological functions and pathways relevant to osteoarthritis.

## Abstract

To construct a diagnostic model of osteoarthritis related to methylation genes using machine learning algorithms, and analyze its prognostic value and biological functions.

The GSE 63695 and GSE162484 datasets including human osteoarthritis (OA) and normal samples were downloaded from the GEO datasets. The microarray chip data of chondrocytes were analyzed using R software to obtain differentially methylated genes. Genes were selected through SVM-RFE analysis and LASSO regression model, and a diagnostic model for OA was established. The performance of the model was assessed by the receiver operating characteristic (ROC) curve. The gene set enrichment analysis of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) was carried out on the genes incorporated within the model.

An overall 11 DEGs were identified:7 genes were remarkably upregulated and 4 genes were distinctly downregulated. By means of machine learning algorithms, ARHGEF10, ATP11A, NOTCH1, THSD4, NIPA1, SIM2, MAN1C1, ENDOG, CCNC, TAF5, and VPS52 were ultimately incorporated into the model, which could effectively diagnose OA. The area under the curve (AUC) in the datasets GSE 63695 and GSE162484 was 0.96 and 0.93 respectively.

The diagnostic model of methylation-related genes constructed based on machine learning algorithms can effectively identify OA.