Methodological pitfalls in the construction of epigenetic clock models

Florent Chuffart
CNRS UMR 5309/INSERM U1209/Université Grenoble-Alpes/Institute for Advanced Biosciences
Presentation of the nature of DNA methylation data captured from Illumina Methylation Assays (Homo sapiens). Evaluation of the impact of cofactors (sex, cell composition, latent factors) on the performance of the machine learning model (elastic-net + booststrap). Illustration: the impact of tobacco consumption on blood's epigenetic age.