Innovative Approaches to Predicting Metabolic Diseases Using Bioinformatics and Machine Learning Technologies: Bioinformatics and Machine Learning Technologies
Year: 2026
Type: article
Abstract: Contemporary approaches to predicting metabolic diseases – such as metabolic syndrome, obesity, and type 2 diabetes – are increasingly grounded in bioinformatics methods, including genomic, metabolomic, and polygenic risk modeling, as well as machine learning and multi-omics data integration. This review highlights key trends in the field: the identification of genetic, transcriptomic, and metabolic biomarkers; the development of polygenic and multi-omics predictive models; the use of cohort and prospective studies; and methodological innovations aimed at improving accuracy and translational relevance. Particular emphasis is placed on machine learning approaches based on genetic and metabolomic data, exploring both their potential and limitations. Recent studies are discussed in which models incorporating genetic variation and dietary factors have achieved high predictive performance. Furthermore, the review examines how integrating “omics” data (genomics and metabolomics) with clinical features enhances prediction, captures inter-population variability, and enables responsiveness to lifestyle factors. Finally, future directions are outlined, including data standardization, increasing ethnic diversity in datasets, development of interpretable and clinically applicable models, and validation of intervention outcomes across independent cohorts.
Institution: Al-Farabi Kazakh National University
Language: English
Subfield: Genetics
Domain: Life Sciences
Open Access status: diamond
OpenAlex ID: W7220145985