BIOLOGICAL AGE: METHODOLOGY FOR DETERMINING

Abstract

The analytical review is devoted to the methodology for quantifying age-related changes, namely, the determination of biological age (BA). In contrast to the chronological age (CA), which is the same for all people born at the same time, BA can vary significantly and characterizes the rate of aging. Diagnostics of accelerated aging is of great clinical importance, as it allows predicting the development of age-related pathology. In addition, the BA calculation is used to assess various effects on the body, including geroprotectors. To calculate BA, a set of biomarkers of aging (BMA) is used. The most accurate method for determining BA is by measuring the methylation of certain genes (Horvath clock). However, this method is not available for clinical practice yet. Therefore, the development of BA models based on clinical BMS remains relevant. In addition, there is a need to assess the degree of age-related changes at the level of individual functional systems of the body (functional age). The list of BMA available for use in clinical practice is given in this review. The requirements are given that allow using the indicator of age-related changes as BMA. The main methods for calculating BV are multiple linear regression (MLR), principal component analysis (PCA), Klemera and Doubal method (KDM), and neural network analysis. The criteria for assessing the quality of the BA model are described in this review. The MLR method, due to its simplicity, is most often used in clinical practice, however, it has a significant drawback, namely, the systematic error in calculating age, which is most pronounced in the extreme age groups. The article provides a method for correcting this error developed by the authors. The methodology of using MLR to obtain a formula for determining BA from a dataset BMA is described in detail. The theoretical foundations of the use of neural network algorithms for assessing the rate of aging are presented. The use of neural networks for assessing BA has shown their high efficiency by all criteria, in particular, the absence of a systematic error inherent in MLR-based methods. Keywords: biological age, biomarkers of aging, calculation methods.