Correlation of characteristic gene and immune infiltration of rheumatoid arthritis based on bioinformatics and machine learning
10.3969/j.issn.1000-484X.2023.12.002
- VernacularTitle:基于生物信息与机器学习分析类风湿关节炎疾病特征基因与免疫细胞的关系
- Author:
Shilei SONG
1
;
Yueping CHEN
;
Feng CHEN
Author Information
1. 广西中医药大学附属瑞康医院,南宁 530011
- Keywords:
Rheumatoid arthritis;
Synovial tissue;
Bioinformatics;
Molecular mechanism;
Immune infiltration;
Machine learning;
Disease characteristic genes
- From:
Chinese Journal of Immunology
2023;39(12):2470-2476
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore pathogenesis,characteristic genes and immune infiltration of rheumatoid arthritis(RA)by bioinformatics and machine learning methods,and to find correlation between characteristic genes and immune cells.Methods:Chips related to RA were obtained from GEO database,gene differences were analyzed by R language,and GO and KEGG were enriched and analyzed;machine learning methods,namely LASSO regression and SVM-RFE were used to screen disease characteristic genes,ROC curve and sample chips were used to detect accuracy of characteristic genes,CIBERPORT algorithm was used to analyze immune infiltration of RA,and correlation between characteristic genes and immune cells was analyzed.Results:A total of 90 differen-tial genes were obtained from GSE12021 and GSE55235,including 64 up-regulated and 26 down-regulated differentially expressed genes.A total of 209 main items were obtained by GO analysis,mainly involving leukocyte activation,lymphocyte activation,B-cell receptor signaling pathway,etc;KEGG analysis showed that chemokine signaling pathway,IL-17 signaling pathway,Toll-like recep-tor signaling pathway and PPAR signaling pathway were closely related to RA;ten key genes such as IGHM,SLAMF8,CXCL10,FNDC4,AIM2,EGR1 and AKR1B10 were screened by mechanical learning method.Disease characteristic genes were IGHM,SLAMF8,CXCL10,AIM2 and AKR1B10 by ROC curve and sample chip;immune infiltration results showed that there were signifi-cant differences between synovial tissues of RA and normal tissues in 9 kinds of immune cells,such as plasma cells,CD8 T cell,CD4 T cell and T cell follicle helper.By analyzing correlation between disease characteristic genes and immune cells,it was found that there were significant correlations between 5 characteristic genes and some immune cells.Conclusion:There is a significant correla-tion between characteristic genes and immune cells of RA,which provides a preliminary basis for follow-up in-depth study of targeted diagnosis and treatment of RA.
