Pilot Study of Item Response Theory-based DNA Methylation Analysis for Aging Association

Abstract

Age-associated DNA methylation, which is commonly observed in multiple tissue and cell types in mammals, is increasingly used as a biomarker for quantifying aging-related health status. In recent years, a set of cytosine-phosphate-guanine (CpG) sites with this DNA methylation, referred to as the epigenetic clock, has been widely used in aging research. This epigenetic clock is considered one of the most promising biomarkers for measuring biological age, which represents physiological functions and capabilities. Though many age-associated CpGs have been found, our understanding of the differences between each CpG’s characteristics and their correlation with aging is still limited. In the current paper, we propose a novel approach that uses Item Response Theory (IRT) to investigate how the DNA methylation profile changes with aging. We found that IRT-based methylation analysis reveals a wide distribution of aging-related features. Most of the CpGs showed specific sensitivity to a particular age range. The results were validated with both whole blood data and CD4+ T cell data sets. Furthermore, we report the possibility of predicting a subject’s potential degree of aging. The predicted degree of aging for subjects had a correlation of 0.8 with chronological age using the whole blood data set. Our pilot study shows that use of an IRT-based analysis for aging association study can expand CpG feature classification.