Can AD be identified through metabolic models?

Abstract

The most common form of dementia is Alzheimer’s Disease (AD), a progressive neurodegenerative disorder, but so far no methods of curing or preventing it are known. Amyloid-β proteins have been the focus of therapeutic development but clinical trials focused on reducing amyloid-β production or preventing its aggregation have failed, leading to increased interests in other areas of AD pathology. There is evidence of impaired metabolism in those with AD, and in this project gene expression data from the ROSMAP study was used to construct single-sample genome-scale metabolic models in an effort to see what differences can be discerned between models from those with AD and not. Although there was no observed separation between the groups when using PCA and t-SNE, the aggregate metabolic coverage between groups differed in several metabolic subsystems. The differences found had support in literature, but subsystems C5-branched dibasic acid metabolism, GPI anchor biosynthesis, heparan sulphate degradation, and lipoic acid metabolism had not been identified as differing in comparable in silico metabolic AD model studies