A distance measure for heterogeneity using genome scale metabolic networks

Physiological differences in the aging process are inherently present in a population, and increase with age, affecting the risk of developing disabilities and age-related diseases [1]. Patient-Derived Genome-Scale Metabolic Models (PD-GSMM) are built from human GSMM and experimental data, mostly transcriptomics and proteomics, belonging to single individuals. Personalized genome scale models have recently been used to plan individualized anti-cancer therapies [2], and to address the variability among cancer patients, identifying key genes involved in tumour growth [3]. Despite their success in cancer metabolism, is still not clear the extent to which PD-GSMs are representations of individual metabolic features in physiological conditions, and how successful such models are in capturing inter-individual heterogeneity when dealing with subtler phenotypes such as ageing. Starting from microarray datasets of younger and older adults’ skeletal muscle gene expression, we developed the first collection of patient-derived genome scale metabolic models of ageing individuals' myocytes, and used a data science approach to define a distance metric and assess the variability between metabolic models. This research is part of the PANINI project (Physical Activity and Nutrition INfluences in Aging), and has received funding from the European Union’s Horizon2020 programme, under the Marie Sklodowska-Curie grant agreement 675003.