Genetics

This post covers Oppong and colleagues' 2024 iScience study using serum NMR metabolomics and whole blood transcriptomics to separate relapsing-remitting from secondary progressive multiple sclerosis. It reports the machine learning performance and what happens when age and EDSS are removed from the models, the points-based score built on five metabolites, the finding that acetoacetate alone outperformed EDSS and age, the direction of the metabolic shift from glycolysis toward gluconeogenesis and ketogenesis, the small sample behind the 1,052 differentially expressed genes, and the treatment-exposure confound between the two patient cohorts that has to be resolved before the signature can become a test.