RESUMO
Bioinformatics and machine learning tools have made it possible to integrate data across different -omics platforms for novel multiomic insights into diseases. To synergistically process -omics data in an integrative manner, analyte extractions for each -omics type need to be done on the same set of clinical samples. Therefore, we introduce a simultaneous dual extraction method for generating both metabolomic (polar metabolites only) and glycomic (protein-derived N-glycans only) profiles from one sample with good extraction efficiency and reproducibility. As proof of the usefulness of the extraction and joint-omics workflow, we applied it on platelet samples obtained from a cohort study comprising 66 coronary heart disease (CHD) patients and 34 matched healthy community-dwelling controls. The metabolomics and N-glycomics data sets were subjected to block partial least-squares-discriminant analysis (block-PLS-DA) based on sparse generalized canonical correlation analysis (CCA) for identifying relevant mechanistic interactions between metabolites and glycans. This joint-omics investigation revealed intermodulative roles that protein-bound carbohydrates or glycoproteins and amino acids have in metabolic pathways and through intermediate protein dysregulations. It also suggested a protective role of the glyco-redox network in CHD, demonstrating proof-of-principle for a joint-omics analysis in providing new insights into disease mechanisms, as enabled by a simultaneous polar metabolite and protein-derived N-glycan extraction workflow.