Kevin Yonghoon Cho (Washington University in St. Louis)

Kevin Yonghoon Cho (Washington University in St. Louis)

Large-scale metabolomics studies require workflows that are robust, scalable, and reproducible. However, large cohorts remain constrained by labor-intensive sample preparation, retention-time drift, and batch effects. We developed an integrated automated platform for sample preparation, LC-MS analysis, and computational processing.

Sample extraction was automated on a Beckman Coulter Biomek i7 using a Captiva EMR-Lipid workflow to sequentially isolate polar metabolites and lipids from a single aliquot. Automation reduced preparation time from more than four hours to approximately 100 minutes per plate while maintaining high reproducibility. Polar metabolites and exposure chemicals were analyzed by high-resolution QTOF mass spectrometry, while more than 700 lipid species were measured by dynamic MRM on an Agilent 6495D triple quadrupole.

An indexed retention-time workflow automatically adjusted peak boundaries across batches, and random forest correction minimized long-term signal drift. Applied to more than 1,800 participants in the NIH-funded Multi-Omics for Health and Disease consortium, the platform produced consistent quality-control performance across multiple batches and reproducibly detected kidney disease-associated metabolites, including dimethylguanidino valerate, glutarylcarnitine, and N2,N2-dimethylguanosine. This workflow provides a scalable framework for metabolomics, lipidomics, and exposomics in large clinical cohorts.