Identification and verification of four candidate biomarkers for early diagnosis of osteoarthritis by machine learning

Xinyu Wang ^1,2, Tianyi Liu ^1,3,4, Yueyang Sheng ¹

⁰Department of Molecular Orthopaedics, National Center for Orthopaedics, Beijing Research Institute of Traumatology and Orthopaedics, Beijing Jishuitan Hospital, Capital Medical University, Beijing, 100035, China.

Heliyon +

|

August 19, 2024

View abstract on PubMed

Summary

This study identified a four-gene biomarker panel (B3GALNT1, GRB10, KLF9, SCRG1) for early osteoarthritis (OA) diagnosis. This discovery offers new pathways for understanding OA mechanisms and developing diagnostic tools.

Area Of Science

Biomarker discovery
Genomics
Osteoarthritis research

Background

Osteoarthritis (OA) is a prevalent chronic joint disease.
Early diagnosis is crucial for effective management.
Identifying reliable diagnostic biomarkers remains a challenge.

Purpose Of The Study

To identify and validate novel diagnostic biomarkers for osteoarthritis.
To assess the clinical significance of candidate biomarkers in patient samples.
To develop a robust diagnostic model for early OA detection.

Main Methods

Utilized three Gene Expression Omnibus (GEO) datasets for training and validation.
Applied machine learning algorithms (Random Forest, LASSO, SVM-RFE) to screen differentially expressed genes (DEGs).
Validated candidate biomarkers using quantitative real-time PCR on peripheral blood samples and chondrocytes.

Main Results

Screened 251 DEGs, with B3GALNT1, SCRG1, and ZNF423 identified by all algorithms.
A combined model including B3GALNT1, GRB10, KLF9, and SCRG1 achieved high AUC values (0.986 training, 1.000 validation, 0.836 test set).
Individual biomarkers showed variable performance, but the combined model demonstrated strong diagnostic potential.

Conclusions

A four-gene signature (B3GALNT1, GRB10, KLF9, SCRG1) was identified as a promising biomarker panel for early osteoarthritis diagnosis.
This panel offers a potential non-invasive diagnostic approach.
Findings provide new insights into OA pathogenesis and potential therapeutic targets.

Keywords:

B3GALNT1 Biomarker Chondrocyte Early diagnosis GRB10 KLF9 Machine learning Osteoarthritis SCRG1 ZNF423