When skeletal muscle mass is lost in ageing and disease, health outcomes can be compromised. Identification of biological signatures that are able to stratify ageing tissue may not only assist to predict those at risk of health decline, but also identify novel preventative measures for a number of age associated diseases. To date, metabolomic approaches have yielded insight into changes in the muscle metabolome with age (1), and how changes in the plasma metabolome relate to age-associated co-morbidities and health (2). What is still not known, is whether changes in the muscle metabolome associated with ageing relate to health, and whether the muscle metabolome could be predictive of ageing. Here we take a unique bioinformatic approach utilising untargeted metabolomics alongside machine learning, in an attempt to identify metabolites associated with chronological ageing and health biomarkers. Fasted muscle biopsies were collected from young (n=10,25±4y), middle-aged (n=18,50±4y), and older (n=18,70±3y) men and women, and untargeted metabolomics data were then generated using UHPLC-MS. Variable importance measures, derived from the Random Forest (RF) algorithm, were used to determine metabolite features most informative in stratifying older age. Potential biological context that may regulate muscle ageing phenotype was examined using the PUIMet algorithm (3). Linear regression was then used to determine any relationship of candidate metabolites to clinical variables commonly associated with age-related health decline. The RF algorithm was able to filter a large dataset of metabolites to those capable of predicting older age. Following putative identification of each metabolite, key subnetworks associated with human muscle ageing included phosphocreatine, androgen metabolism, histamine and phospholipid metabolism. Following FDR corrected linear regression, the phospholipid metabolite LysoPE(20:5) showed a significant negative relationship with leg lean mass (LLM), strength and resting heart rate (RHR), and LysoPE(22:6) showed a positive relationship to %Body Fat. Expected relationships were observed between dehydroepiandrosterone sulfate and total cholesterol, and the fatty acid metabolite capryloylglycine with HDL. However, others such as spermine with HDL and RHR, histamine with mean arterial pressure (MAP), dihydrothymine with insulin, and 4-imidazolone-5-proprionic acid with HOMA-IR, showed potential novel metabolite links to age associated disease factors. This study reveals that a novel bioinformatics based metabolomics approach is potentially useful to classify human tissue ageing, and intriguingly, that a number of the age predictive metabolites show associations with common health biomarkers, and therefore could be used to assist towards the identification of novel preventative measures for a number of age associated diseases.