TY - JOUR T1 - DNA methylation biomarker in blood predicts frailty in an HIV-positive veteran population JF - medRxiv DO - 10.1101/19010272 SP - 19010272 AU - Chang Shu AU - Amy C. Justice AU - Xinyu Zhang AU - Vincent C. Marconi AU - Dana B. Hancock AU - Eric O. Johnson AU - Ke Xu Y1 - 2019/01/01 UR - http://medrxiv.org/content/early/2019/10/29/19010272.abstract N2 - Background With the improved life expectancy in people living with HIV, predicting individual frailty is important for clinical care. DNA methylation (DNAm) has emerged as a robust biomarker in precision medicine and has been previously linked to aging and mortality in non-HIV populations. Here, we aim to establish a panel of DNAm features selected from blood methylome to predict frailty in HIV-positive men from a veteran population, Veterans Aging Cohort Study (VACS).Methods We used a well-established score, VACS Index, as a measure of frailty. Samples (ntotal=1,150) were divided into a training set (n=612) and a validation set (n=538). We first selected a panel of frailty-associated CpGs by conducting an epigenome-wide association analysis on the VACS index in the training set. We then applied four machine learning methods to build models to predict high and low frailty individuals in the training set. The prediction models were tested in the validation set. A methylation score constructed from the selected CpGs was tested an associated with mortality by performing survival analysis. To assess the biological relevance of the selected CpG sites, we performed a gene ontology enrichment analysis.Results A panel of 119 CpGs were identified from the training set (False Discovery Rate <0.2) that showed excellent performance on predicting high frailty individuals with Area Under Curve (AUC) of 0.835 (95%CI: 0.792, 0.879) and balanced accuracy of 0.693. The same panel showed good performance on predicting low frailty individuals with AUC of 0.735 (95%CI: 0.688-0.782) and a balanced accuracy of 0.528. The methylation score from the 119 CpGs was significantly associated with 5-year and 10-year mortality with hazard ratio of 1.40 (95% CI:1.033, 1.897 p=0.03) and 1.48 (95%CI: 1.10, 2.02; p=0.01) respectively. These 119 CpGs were located within or near 73 genes that were significantly enriched in 9 biological pathways relevant to immune and inflammation response.Conclusions We identified a panel of predictive DNAm features associated with frailty and mortality among HIV-positive population. These DNAm features may serve as biomarkers to discriminate high and low frailty people who live with HIV.Competing Interest StatementThe authors have declared no competing interest.Funding StatementNational Institute on Drug AbuseAuthor DeclarationsAll relevant ethical guidelines have been followed and any necessary IRB and/or ethics committee approvals have been obtained.YesAll necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesAny clinical trials involved have been registered with an ICMJE-approved registry such as ClinicalTrials.gov and the trial ID is included in the manuscript.Not ApplicableI have followed all appropriate research reporting guidelines and uploaded the relevant Equator, ICMJE or other checklist(s) as supplementary files, if applicable.YesData has been deposited in GEOAUCArea Under CurveCIConfidence intervalDMRdifferentially methylated regionDNADeoxyribonucleic acidDNAmDNA methylationDAVIDDatabase for Annotation, Visualization and Integrated DiscoveryEWASepigenome-wide association studyFDRFalse discovery rateFWERFamily-wise error rateGLMNETLasso and Elastic-Net Regularized Generalized Linear ModelsGOGene ontologyHIVHuman immunodeficiency virusHM450KHuman Methylation 450K BeadChipNKNatural killerPCPrincipal componentQCQuality controlSVMSupport Vector MachinesVACSVeterans Aging Cohort StudyVACS indexVeterans Aging Cohort Study indexXGBoostExtreme Gradient Boosting Tree ER -