Building machine learning prediction models for wellbeing using predictors from the exposome and genome in a population cohort

Dirk Pelt, Philippe Habets, Christiaan H. Vinkers, Lannie Ligthart, CEM van Beijsterveldt, René Pool, Meike Bartels

Research output: Working paper / PreprintPreprintAcademic

Abstract

Using longitudinal data of a large population cohort (Netherlands Twin Register; collected 1991-2022), we aim to build machine learning prediction models for adult wellbeing from the exposome and genome, and identify the most predictive factors (Ns between 702 and 5874). The specific exposome was captured by parent- and self-reports of psychosocial factors from childhood to adulthood, the genome by polygenic scores, and the general exposome by linkage of participants’ postal codes to objective exposures based on registry data. The specific exposome was found to be highly predictive of wellbeing in an independent test set (R2 = .701), but not the genome (R2 = .008) and general exposome (R2 = -.006) independently, nor incrementally. Risk/protective factors such as optimism, personality, social support, and previous wellbeing levels were most predictive. Our findings highlight the importance of monitoring individuals over time and current limitations associated with prediction based on the genome and general exposome.
Original languageEnglish
DOIs
Publication statusSubmitted - 13 Oct 2023

Fingerprint

Dive into the research topics of 'Building machine learning prediction models for wellbeing using predictors from the exposome and genome in a population cohort'. Together they form a unique fingerprint.

Cite this