Vertical Metabolome Transfer from Mother to Child: An Explainable Machine Learning Method for Detecting Metabolomic Heritability.

ABSTRACT: Vertical transmission of metabolic constituents from mother to child contributes to the manifestation of disease phenotypes in early life. This study probes the vertical transmission of metabolites from mothers to offspring by utilizing machine learning techniques to differentiate between true mother-child dyads and randomly paired non-dyads. Employing random forests (RF), light gradient boosting machine (LGBM), and logistic regression (Elasticnet) models, we analyzed metabolite concentration discrepancies in mother-child pairs, with maternal plasma sampled at 24 weeks of gestation and children's plasma at 6 months. The propensity of vertical transfer was quantified, reflecting the likelihood of accurate mother-child matching. Our findings were substantiated against an external test set an

SUBMITTER: Lovric M 

PROVIDER: S-EPMC10972480 | biostudies-literature | 2024 Feb

REPOSITORIES: biostudies-literature