Machine learning for screening active metabolites with metabolomics in environmental science

Abstract

Metabolites are substances produced during metabolism, playing roles in biological processes such as biochemical reactions, signaling and gene expression. Metabolomics is a study of all metabolites and metabolic patterns in the body in response to genetic and environmental stresses. With advanced analytical techniques, metabolomic analysis generates large-scale and complex datasets that must be interpreted for meaningful biological information. Machine learning is an emerging area and applied to reveal the data structure, achieve the predictability of trends, and discover the metabolic patterns of metabolomic data in environmental science. Here, we review the applicability of machine learning to screen active metabolites with metabolomics in environmental science, while presenting the use of machine learning for metabolomics data processing, toxic effects of environmental pollutants, and health outcomes of environmental exposure. We also discuss the potential of combining integrative metabolomics with novel machine learning algorithms for the challenges of complex relationships between active metabolites and environmental exposures.