Screening and mechanism research on pyroptosis-related genes in atrial fibrillation based on machine learning

Wang Qiuyu; Tan Pengjin; Li Xin; Shang Lihua; Liu Jianguo

doi:10.3969/j.issn.1674-4055.2024.10.05

Abstract：Objective To investigate the effect and potential mechanism of pyroptosis in occurrence and development of atrial fibrillation(AF),and to screen potential therapeutical targets.Methods The gene expression omnibus(GEO)was retrieved for downloading expression profile data and clinical information of AF.The batch effects among different GEO datasets were removed by using ComBat algorithm,and integrated as merge-cohort.Subsequently,the expression data of 32 pyroptosis-related genes were extracted for differential analysis by using student T-test.AF risk score model and gene set were established by using single-factor Logistic regression analysis and LASSO regression analysis,and the potential mechanism of AF occurrence and development was explored by using enrichment analysis.The pyroptosis-related genes,which played a key role in AF development,was screened by using XG-boost machine learning algorithm,and diagnostic value of pyroptosis-related genes was analyzed by using ROC curve analysis.Results There were significant expression differences among pyroptosis-related genes in AF and sinus rhythm atrial tissues.A 5-gene risk score model was established taken integrated merge-cohort as object(risk score=0.734 × PYCARD+0.435 × SCAF11+0.008 × CASP8-0.295 × GSDMC-0.342 × IL 18).The results of enrichment analysis showed that pyroptosis was involved AF occurrence and development through regulating inflammatory infiltration.The results of machine learning showed that PYCARD played a key role in AF occurrence and development with a good diagnostic value.Conclusion Pyroptosis has an important influence on AF occurrence and development,and PYCARD,a pyroptosis-related gene,may serve as a new prognostic biomarker and potential therapeutic target for AF.