https://www.biorxiv.org/content/early/2018/02/13/264374.1
https://pubs.acs.org/doi/10.1021/acs.jproteome.9b00319
problem and objectives
The link between abnormal fetal metabolomic profiles and maternal obesity has not yet been fully elucidated. In this study, we analyzed the fetal cord blood metabolomes of obese and non-obese mothers to identify potential metabolomic biomarkers associated with maternal obesity.methods
A penalized regression model was used to identify metabolites which contributed the most to classifying obese and non-obese samples. We adjusted metabolites for ethnicity, gravidity, parity and maternal age. Consensus Pathway database (CPDB) online tool was used for metabolomics pathway analysis.
results
Of 230 detected untargeted and targeted metabolites, 29 were selected by the elastic net penalized regression method. The accuracy of the logistic regression model using these 29 metabolites (obese versus normal) trained on 80% of the samples and tested on 20% was 0.97. In the obese samples, 17 metabolites were increased, and 12 were decreased. Galactonic acid was identified as the most important predictor of obese-normal status in our cohort.
