Establishing Human Age Prediction Model Based On DNA Methylation Characteristics

Aging is a natural and gradual process in human life,affected by many factors such as heredity and environment.Finding reliable biomarkers of aging and methods to delay physiological aging have always been one of the fields that scientists are keen to study.Disorder of epigenetic regulation is an important feature of aging.During the aging process,changes in epigenetic information patterns within each cell in a population will lead to transcription drift and genomic instability.Compared with the genetic regulation represented by DNA sequences,most epigenetic modifications are reversible.Therefore,describing and understanding the epigenetic changes that occur during aging is an important research direction.Epigenetics has broad prospects for targeted intervention and treatment of related diseases through external conditions.Through the study of epigenetic changes in the aging process,we can also better understand how to delay aging through external means.DNA methylation is a kind of epigenetic marker that can be quantitatively detected in human population,and the change of its state is an important factor affecting aging.Recent studies have proved that the whole genome DNA methylation of human decreases significantly with age.Many studies have reported that specific methylation modification sites have a significant correlation with age.At present,there are methods to predict the degree of human aging based on DNA methylation sites.In many cases,one site may target multiple genes,or multiple sites may target a common gene.In addition,methylation modification has a certain degree on the genome.Because of the randomness,the site cannot directly reflect the methylation status of phenotype-related genes.As a basic unit of biological function,it is not clear whether genome-wide changes in gene methylation levels are similar to those at aging sites,or whether methylation genes can be directly analyzed and an age prediction model can be established.This study used the methylation detection data from the GEO database,the SIMPO(Statistical difference of Methylation between Promoter and Other Body Region)algorithm was used to calculate the methylation characteristic values of genes,then use a variety of regression methods to explore,and finally establish an age prediction model containing 71 methylated genes.After completing the establishment and optimization of the model,the independent test data set was used to test the model and compared with the published age prediction model based on methylation sites,and the age prediction error of the model based on methylation genes is 5.5 years,which is comparable to traditional models based on methylation sites.In addition,the gene function enrichment analysis results show that the model established in this study is more closely related to aging and development and regeneration,and has more significant biological significance.On the other hand,the results of the Parkinson's disease dataset test and the gene function analysis based on the Knowledge Graph also show that the age prediction model established in this study is closely related to known aging-related diseases,this suggests that starting from the apparent regulation of methylation may help prevent aging-related diseases.