This repository contains code for the final project of Computational Biology of Aging course (cohort 2024).

Quantifying the stochastic component of epigenetic aging

Category: reproduction of the paper

Short introduction

Epigenetic clocks are multivariate linear models that can predict chronological and biological age, based on DNA methylation (DNAm) data. Age-associated DNAm changes are observed in two phenomena – first, the evidence of specific age-associated DNAm changes, and second, the evidence of erosion or increasing uniformity of DNAm landscape, associated with age. Thus, two phenomena can be considered deterministic and stochastic components of quasi-stochastic epigenetic changes associated with age. In this regard, understanding the contribution of stochastic component to the accuracy of linear models (clocks) prediction represents a fundamental interest.

Main hypothesis:

According to the paper, approximately 66–75% of the accuracy underpinning Horvath’s clock could be driven by a stochastic process.

Goals:

1. Reproduce the stochastic simulation for Horvath’s clock;
2. Construct a stochastic analog (StocH clock) using DNAm dataset;
3. Quantify the stochastic component of Horvath’s clock in one of the sorted immune cell datasets.

Data availability: GSE56581, GSE56046
Related papers: Tong, H., Dwaraka, V.B., Chen, Q. et al. Quantifying the stochastic component of epigenetic aging. Nat Aging 4, 886–901 (2024).
Responsible TA: Dmitrii Kriukov

Team: Oksana Kotovskaya