Diagnostics and association of Alzheimer’s disease markers with metabolomic age.

Every cellular and metabolic process slows down with either regular chronological aging or induced accelerated aging by a chronic condition such as AD. In this aim, I propose to understand how the metabolites from cognitively normal subjects differ from the metabolites in AD subjects with a predicted model to calculate the brain metabolic age. To further validate my research, I plan on associating the predicted brain metabolic age to phenotypic data including MRI phenotypes, cognitive scores, plasma and CSF biomarkers. The bioenergetic rate of neurodegenerative brains have been reported to slow down at higher pace than regular brains mainly due to friction of oxidative and metabolic stress. My research wanted to understand how this metabolic stress can be used as a biomarker to AD progression by understanding how and which AD phenotypes the metabolic age associates with the best including MRI imaging phenotype, cognitive scores, CSF and plasma biomarkers. The significance of this study relies on training our model to predict the metabolic age on a dataset completely made of heathy subjects and predicting the metabolic age of subjects with AD on a different dataset. Such data crossover studies provide more insight on common factors of AD pathology. Furthermore, with longitudinal data and sex-based training, we hope to provide some insights on importance of metabolic changes in AD progression.

Our pipeline of utilizing the healthy subjects of one dataset to investigate the AD subjects and along with the longitudinal effect of the same metabolic is one of a kind. Our novelty stands in our process with also training the model using all the metabolites and how well it transfers over to the AD dataset. We also performed one-of-a-kind study with validation the multi-omics biomarker phenotypes of AD in terms of imaging, cognitive, CSF and plasma with the same metabolic age with the covariates. This study can be set as a foundation for better exploration of the bioenergetics as biomarker from the metabolome and insights from this study can be used for early detection of AD with just the metabolomic data which can be extracted much easier compared to the expression sample of the brain.