Journal article
Data integration for prediction of weight loss in randomized controlled dietary trials
Department of Health Technology, Technical University of Denmark1
Harvard University2
Bispebjerg University Hospital3
Department of Biotechnology and Biomedicine, Technical University of Denmark4
Disease Systems Immunology, Section for Protein Science and Biotherapeutics, Department of Biotechnology and Biomedicine, Technical University of Denmark5
Bioinformatics, Department of Health Technology, Technical University of Denmark6
Disease Data Intelligence, Bioinformatics, Department of Health Technology, Technical University of Denmark7
National Food Institute, Technical University of Denmark8
Research Group for Genomic Epidemiology, National Food Institute, Technical University of Denmark9
University of Copenhagen10
T-Cells and Cancer, Experimental & Translational Immunology, Department of Health Technology, Technical University of Denmark11
Research Group for Gut, Microbes and Health, National Food Institute, Technical University of Denmark12
Technical University of Denmark13
...and 3 moreDiet is an important component in weight management strategies, but heterogeneous responses to the same diet make it difficult to foresee individual weight-loss outcomes. Omics-based technologies now allow for analysis of multiple factors for weight loss prediction at the individual level. Here, we classify weight loss responders (N = 106) and non-responders (N = 97) of overweight non-diabetic middle-aged Danes to two earlier reported dietary trials over 8 weeks.
Random forest models integrated gut microbiome, host genetics, urine metabolome, measures of physiology and anthropometrics measured prior to any dietary intervention to identify individual predisposing features of weight loss in combination with diet. The most predictive models for weight loss included features of diet, gut bacterial species and urine metabolites (ROC-AUC: 0.84–0.88) compared to a diet-only model (ROC-AUC: 0.62).
A model ensemble integrating multi-omics identified 64% of the non-responders with 80% confidence. Such models will be useful to assist in selecting appropriate weight management strategies, as individual predisposition to diet response varies.
Language: | English |
---|---|
Publisher: | Nature Publishing Group UK |
Year: | 2020 |
Pages: | 20103 |
ISSN: | 20452322 |
Types: | Journal article |
DOI: | 10.1038/s41598-020-76097-z |
ORCIDs: | Nielsen, Rikke Linnemann , Helenius, Marianne , Garcia, Sara L. , Roager, Henrik M. , Aytan-Aktug, Derya , Bahl, Martin I. , Brix, Susanne , Petersen, Thomas Nordahl , Licht, Tine Rask , Gupta, Ramneek , 0000-0002-4999-1218 , 0000-0002-0065-8174 , 0000-0001-8748-3831 , 0000-0002-6024-0917 , 0000-0001-7184-5949 and 0000-0002-3321-3972 |
Biomarkers Computational models Computer modelling Computer science DNA Data integration Data processing Diet Therapy Female Gastrointestinal Microbiome Genome-Wide Association Study Genotype Humans Machine Learning Machine learning Male Metabolomics Microbial genetics Postprandial Period Predictive medicine ROC Curve Randomized Controlled Trials as Topic Reproducibility of Results Sequencing Treatment Outcome Weight Loss Weight management Whole Grains