Clinical Validation of Targeted and Untargeted Metabolomics Testing for Genetic Disorders: A 3 Year Comparative Study

• Published in: Scientific reports Vol. 10; no. 1; p. 9382
• Main Authors: Almontashiri, Naif A. M., Zha, Li, Young, Kim, Law, Terence, Kellogg, Mark D., Bodamer, Olaf A., Peake, Roy W. A.
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 10.06.2020; Nature Publishing Group
• Subjects: 631/45; 631/45/320; 692/53; 692/53/2421; Adolescent; Biomarkers - metabolism; Child; Child, Preschool; Cohort Studies; Comparative studies; Diagnosis; Female; Genetic Diseases, Inborn - genetics; Genetic disorders; Genetic Testing; Humanities and Social Sciences; Humans; Inborn errors of metabolism; Male; Metabolism, Inborn Errors - genetics; Metabolites; Metabolomics; Metabolomics - methods; multidisciplinary; Phenotype; Phenotypes; Science; Science (multidisciplinary); Syndrome
• DOI: 10.1038/s41598-020-66401-2
• ISSN: 2045-2322

Global untargeted metabolomics (GUM) has entered clinical diagnostics for genetic disorders. We compared the clinical utility of GUM with traditional targeted metabolomics (TM) as a screening tool in patients with established genetic disorders and determined the scope of GUM as a discovery tool in patients with no diagnosis under investigation. We compared TM and GUM data in 226 patients. The first cohort (n = 87) included patients with confirmed inborn errors of metabolism (IEM) and genetic syndromes; the second cohort (n = 139) included patients without diagnosis who were undergoing evaluation for a genetic disorder. In patients with known disorders (n = 87), GUM performed with a sensitivity of 86% (95% CI: 78–91) compared with TM for the detection of 51 diagnostic metabolites. The diagnostic yield of GUM in patients under evaluation with no established diagnosis (n = 139) was 0.7%. GUM successfully detected the majority of diagnostic compounds associated with known IEMs. The diagnostic yield of both targeted and untargeted metabolomics studies is low when assessing patients with non-specific, neurological phenotypes. GUM shows promise as a validation tool for variants of unknown significance in candidate genes in patients with non-specific phenotypes.