Integration of clinical metabolomic profiling with interpretation of genetic variants improves diagnosis of rare disease
Sarah H. Elsea, Ph.D. FACMG
Senior Division Director, Biochemical Genetics
Dept. of Molecular and Human Genetics
Baylor College of Medicine
Houston, TX USA
Metabolomics is the study of the distinctive chemical fingerprint produced by specific cellular processes. Untargeted mass spectrometry-based metabolomics profiling for small molecules in body fluids, including urine, plasma, and cerebrospinal fluid, is an emerging technique used to produce and analyze this chemical fingerprint. This technology holds the promise of providing new insights into human disease states and serving as a primary diagnostic tool for novel and previously characterized inborn errors of metabolism (IEM), as well as for the identification of biomarkers of disease and treatment. Clinical metabolomic profiling is a novel platform that allows for parallel testing of hundreds of metabolites in a single biological specimen. Using a state-of-the-art mass spectrometry platform, the resulting spectra are compared against a library of ~2,500 human metabolites. On average, ~800 small molecules are detected in a given plasma sample with a core group of ~350 analytes found in all specimens tested to date. The analytes detected encompass numerous classes of important small molecule biomarkers including acylcarnitines, amino acids, bile acids, carbohydrates, lipids, and nucleotides. In addition, metabolomic data in many cases affords a much richer view of a patient’s metabolic disturbance by identifying: (1) elevated metabolites located far upstream of the genetic defect, (2) treatment related compounds, including commonly tested therapeutic drug monitoring analytes, and (3) spectrally unique analytes that are not yet associated with a biochemical phenotype. In our clinical experience, the integration of whole exome sequencing data with the metabolomics profile has improved the interpretation of genetic variants, including ruling out the diagnosis of IEMs, as well as supporting a specific diagnosis, and for the identification of new disease biomarkers. For the undifferentiated genetic phenotypes such as intellectual disability, autism, or seizures, often many different tests involving different sample types are needed for diagnosis. This can lead to prohibitive costs and ongoing diagnostic odysseys. Utilizing clinical metabolomics alone as a broad screening tool has also resulted in several diagnoses, typically expanding the phenotypes of those disorders. Our experience with metabolomic profiling of previously nondiagnostic cases has led to the diagnosis of genetic disorders such as GABA transaminase deficiency, aromatic amino acid decarboxylase deficiency, and adenylosuccinate lyase deficiency, illustrating the powerful synergy of metabolomic and genomic analyses in determining the pathogenicity of variants of uncertain significance, as well as broadening the phenotypic spectrum of each disorder. Ultimately, a clinical systems biology approach to the integration clinical data with genomic, transcriptomic, epigenomic, proteomic, and metabolomics data will provide a better understanding of natural biological variation toward improved diagnosis.