Metabolic network changes that are strongly associated with Dementia with Lewy Bodies as determined through signed distance and partial correlation analysis

MotivationIdentifying pathological metabolic changes in complex disease such as Dementia with Lewy Bodies (DLB) requires a deep understanding of functional modifications in the context of metabolic networks. Network determination and analysis from metabolomics and lipidomics data remains a major challenge due to sparse experimental coverage, a variety of different functional relationships between metabolites and lipids, and only sporadically described reaction networks. ResultsDistance correlation, measuring linear and non-linear dependences between variables as well as correlation between vectors of different lengths, e.g. different sample sizes, is presented as an approach for data-driven metabolic network development. Additionally, novel approaches for the analysis of changes in pair-wise correlation as well as overall correlations for metabolites in different conditions are introduced and demonstrated on DLB data. Distance correlation and signed distance correlation was utilized to determine metabolic network in brain in DLB patients and matching controls and results for the two groups are compared in order to identify metabolites with the largest functional change in their network in the disease state. Novel correlation network analysis showed alterations in the metabolic network in DLB brains relative to the controls, with the largest differences observed in O-phosphocholine, fructose, propylene-glycol, pantothenate, thereby providing novel insights into DLB pathology only made apparent through network investigation with presented methods.