Current Advancements in Untargeted Metabolomics Analysis and Testing Driven by Machine Learning: Prospects for...

Metabolomics is a high-throughput methodology that measures various metabolites in biological fluids. An advantage of an untargeted approach for analyzing metabolomics, which involves an impartial examination of the metabolome, lies in its ability to identify crucial metabolites that contribute to, or serve as indicators of, human health and disease. This review discusses the significance of artificial intelligence and machine learning in advancing disease diagnosis and small molecule detection using untargeted metabolomics, providing an updated and integrative perspective on their recent impact in this field. The discussion focuses on how artificial intelligence and machine learning have contributed to the evolution of high-resolution mass spectrometry. This approach involves unbiased detection of endogenous and exogenous biochemical compounds and metabolites in biological tissue. It enables the characterization of exposures linked to disease outcomes by providing context for the applications of artificial intelligence and machine learning and exploring their roles in enhancing the precision and efficiency of disease exposure assessments. While several review articles address the application of artificial intelligence and machine learning in metabolomics, this article aims to highlight recent opportunities for leveraging these technologies specifically in untargeted metabolomics, covering improvements in data quality, methodological rigor, sensitivity in detection, and precise compound identification. The novelty holds transformative potential that artificial intelligence and machine learning bring to metabolomics, paving the way for more robust and insightful investigations.