IntroductionOnly 40% of mammalian genes have a known function, yet many are associated with human disease. The International Mouse Phenotyping Consortium (IMPC) intends to knockout and phenotype all 20,000 protein‐coding genes in mice to determine new functions of mammalian genes. In this study, we performed metabolomics on 30 knockout mice to determine the utility of metabolomics in linking the mechanisms between genes and disease phenotypes.MethodsGenes were knocked out with conditional gene targeting in embryonic stem cells, which generated homozygote transgenic C57BL6N mice. At 16 weeks of age, mice were terminated, and blood plasma was collected and immediately frozen. Mouse plasma from 30 knockouts (n=6/genotype, male and female) with wild‐type controls (n=40, female and male) were analyzed at the West Coast Metabolomics Center (WCMC) with 6 metabolomics platforms: complex lipids (CSH (+/−) ESI), biogenic amines (HILIC), primary metabolites (GC‐TOF), steroids and bile acids (QTRAP), and oxylipins (QTRAP). 797 unique metabolites were identified and used to compare knockouts to wild‐type using univariate, multivariate and enrichment statistical analyses.ResultsEvery mouse knockout had significant metabolic changes compared to the wild‐type. Lmbrd1, a lysosomal gene involved in vitamin B12 metabolism, had the strongest metabolic phenotype with 166 significantly changed metabolites and 16 significantly altered metabolite clusters. Hierarchal clustering of mouse knockouts by metabolic phenotype demonstrated that genes of similar functions cluster together. ATP6vd01 and ATP5b, both of which code for proteins of ATP synthase, cluster closely together and are characterized by a decrease in several lipid classes. We investigated the metabolic profile of individual genes, such as the well characterized TCA cycle gene Idh1, isocitrate dehydrogenase. We measured expected changes in TCA cycle intermediates: isocitrate was increased (p = 0.0008, fold change = 1.58), whereas alpha‐ketoglutarate was decreased (p = 0.01, fold change = 0.10). Unexpectedly, Idh1 (−/−) mice had lower plasma level of phosphatidylethanolamines (p = 0.01), phosphatidylcholines (p = 6.2E‐05), and phosphatidylinositols (p = 1.5E‐05). In addition to Idh1, we discovered that Mfap4, a gene that is deleted in Smith‐Magenis Syndrome (SMS), results in lipid dysregulation, which may contribute to the severe developmental phenotype of SMS patients. Sra1, a gene that is overexpressed in breast and ovarian cancer, has drastic effects on the lipid profile in female mice, indicating that the tumor‐potentiating effects of this gene may be hormone‐linked.DiscussionOur results demonstrate that metabolomics can be used to determine new metabolic functions of genes with unknown functions, well‐known metabolic genes, and genes with no known metabolic associations. The metabolic phenotype of the mouse knockouts can be linked to precision medicine, aiding in the development of targeted therapies for genes involved in disease susceptibility or progression. Metabolomics should be included in the IMPC phenotyping pipeline to mechanistically link genes to disease phenotypes.This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.