Seminario: Leo Cheng - Lunes 29 de junio 11:00 hs

NMR Metabolomics & NMR-metabolomics-AI-guided Trans-omics Studies of Human Diseases

NMR as a fundamental tool for untargeted and targeted metabolomics is valued for its reproducibility, quantitative accuracy, and broad applicability across various scientific fields. For medical research and application, particularly using our previously invented high-resolution magic angle spinning (HRMAS) technique, NMR metabolomics can measure untreated pathological tissues prior to their pathological evaluations, biological fluids (serum, CSF, etc.) in 10uL scales, and live cells in the amount of <100,000. Furthermore, NMR metabolomics results can be quantitatively correlated with other omics’ results measured from other platforms. In this presentation, we will illustrate the super utilities of NMR metabolomics in medicine, with a primary focus on discussions of our NMR-metabolomics-AI-guided trans-omics evaluations of Human Acute Respiratory Distress Syndrome (ARDS), a severe, life-threatening clinical condition and representing ~10% of all ICU patients in the US, and with its molecular and metabolic mechanisms remain incompletely understood.

To address this gap, we used HRMAS NMR to investigate serum metabolomic profiles in ICU patients with and without ARDS. These metabolomic data were integrated with detailed clinical outcomes and serum proteomics measured using the Olink Proteomics platform. Olink technology combines proximity extension assay (PEA) with next-generation sequencing (NGS), enabling sensitive and high-throughput protein quantification and providing insight into biological processes underlying disease progression.
NMR analysis of pre–COVID-19 serum samples, 74 ARDS patients (64 with matched Olink proteomics data) and 100 non-ARDS ICU control identified 37 spectral-region-of-interest (ROI) and various ROIs can differentiate ARDS from non-ARDS, and correlate with ARDS clinical criteria of the 28-day patient outcome. Olink analyses identified 8 FDR significant proteins. Quantitative analyses between NMR ROIs and Olink proteins identified correlations between150 proteins and 14 ROIs, with identifications of 104 potential metabolites.
Proteomics-metabolomics pathway connections between 150-proteins and 104-metabolites were investigated using Claude AI, that identified 8 potential pathways involving 21 unique proteins and 14 possible metabolites, that can improve protein ROC AUC from 0.89 to 0.92, and NMR AUC from 0.70 to 0.78 for differentiating patient 28-day outcomes.