Plasma metabolomic profiles enhance precision medicine for volunteers of normal health - PubMed (original) (raw)

. 2015 Sep 1;112(35):E4901-10.

doi: 10.1073/pnas.1508425112. Epub 2015 Aug 17.

Michael V Milburn 2, John A Ryals 2, Shaun C Lonergan 2, Matthew W Mitchell 2, Jacob E Wulff 2, Danny C Alexander 2, Anne M Evans 2, Brandi Bridgewater 2, Luke Miller 2, Manuel L Gonzalez-Garay 3, C Thomas Caskey 4

Affiliations

Plasma metabolomic profiles enhance precision medicine for volunteers of normal health

Lining Guo et al. Proc Natl Acad Sci U S A. 2015.

Abstract

Precision medicine, taking account of human individuality in genes, environment, and lifestyle for early disease diagnosis and individualized therapy, has shown great promise to transform medical care. Nontargeted metabolomics, with the ability to detect broad classes of biochemicals, can provide a comprehensive functional phenotype integrating clinical phenotypes with genetic and nongenetic factors. To test the application of metabolomics in individual diagnosis, we conducted a metabolomics analysis on plasma samples collected from 80 volunteers of normal health with complete medical records and three-generation pedigrees. Using a broad-spectrum metabolomics platform consisting of liquid chromatography and GC coupled with MS, we profiled nearly 600 metabolites covering 72 biochemical pathways in all major branches of biosynthesis, catabolism, gut microbiome activities, and xenobiotics. Statistical analysis revealed a considerable range of variation and potential metabolic abnormalities across the individuals in this cohort. Examination of the convergence of metabolomics profiles with whole-exon sequences (WESs) provided an effective approach to assess and interpret clinical significance of genetic mutations, as shown in a number of cases, including fructose intolerance, xanthinuria, and carnitine deficiency. Metabolic abnormalities consistent with early indications of diabetes, liver dysfunction, and disruption of gut microbiome homeostasis were identified in several volunteers. Additionally, diverse metabolic responses to medications among the volunteers may assist to identify therapeutic effects and sensitivity to toxicity. The results of this study demonstrate that metabolomics could be an effective approach to complement next generation sequencing (NGS) for disease risk analysis, disease monitoring, and drug management in our goal toward precision care.

Keywords: disease assessment; functional phenotyping; gene penetrance; metabolomics; whole-exome sequencing.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Fig. 1.

Classification of the plasma metabolites detected in this study. (A) Total number of metabolites based on their biochemical classes. (B) Metabolites modulated by gut bacteria activities. Nonitalicized metabolites are exclusively or mainly contributed by bacteria metabolism, and italicized metabolites are jointly contributed by both mammalian cells and bacteria.

Fig. 2.

Fig. 2.

Metabolomics diversity of the cohort illustrated by the heat map of the metabolomic profiles of the volunteers. Red and blue indicate high and low levels, respectively, relative to the median value for all samples (median = 1.0). The workflow to generate new candidate genes from metabolomics data is shown in Fig. 3. The vCard files are annotated using spnEff and ANNOVAR. Nonsynonymous coding variants are identified. A list of genes corresponding to an abnormal metabolic pathway is generated, and nonsynonymous coding variants are isolated. Variants are filtered using frequency and functional effect filters. Details of the process and methods are provided in Materials and Methods.

Fig. 3.

Fig. 3.

Work flow for searching WES data and metabolomics convergence. MAF, minor allele frequency.

Fig. 4.

Fig. 4.

Assessment of the metabolic perturbations defined biochemical pathways. (A) Purine degradation pathway and dot plots showing data distribution in the cohort for xanthine, urate, and hypoxanthine. The red dots show the metabolite level for volunteer 3923. The open dots show the data distribution for the rest of the cohort (n = 80). (B) Sorbitol degradation pathway and dot plots showing data distribution in the cohort for fructose and sorbitol. The red dots show the metabolite level for volunteer 3905. The open dots show the data distribution for the rest of the cohort (n = 80). The box represents the middle 50% of the distribution, and left and right ‘‘whiskers’’ represent the entire spread of the data. The vertical line refers to the median, and the plus symbol refers to the mean. The first and second numbers within the parentheses are the z-score and P value, respectively.

Fig. 5.

Fig. 5.

Data distribution for long-chain fatty acid carnitines and 3-methylhistidine in the cohort (n = 80). The red dots show the metabolite level for volunteer 3890. The open dots show the data distribution for the rest of the cohort. An explanation of the plots is provided in the legend for Fig. 4. The first and second numbers within the parentheses are the z-score and P value, respectively.

Fig. 6.

Fig. 6.

Bile acid circulation and dot plots showing data distribution in the cohort (n = 80) for the four primary bile acids. The red dots show the metabolite level for either volunteer 3917 or volunteer 3952. The open dots show the data distribution for the rest of the cohort. The box represents the middle 50% of the distribution, and left and right ‘‘whiskers’’ represent the entire spread of the data. The vertical line refers to the median, and the plus refers to the mean. An explanation of the plots is provided in the legend for Fig. 4. The first and second numbers within the parentheses are the z-score and P value, respectively.

Fig. 7.

Fig. 7.

Condensed metabolic schemes for energy metabolism and dot plots showing data distribution in the cohort (n = 80) for key metabolites with a known association with diabetes. The red dots show the metabolite level for the specific volunteers as labeled next to the plots. The open dots show the data distribution for the rest of the cohort. An explanation of the plots is provided in the legend for Fig. 4. The first and second numbers within the parentheses are the z-score and P value, respectively.

Fig. 8.

Fig. 8.

Distribution of key metabolites with known association with diabetes in the cohort (n = 80). The red dots show the metabolite level for either volunteer 3891 or volunteer 3837. The open dots show the data distribution for the rest of the cohort. An explanation of the plots is provided in the legend for Fig. 4. The first and second numbers within the parenthesis are the z-score and P value, respectively.

Fig. 9.

Fig. 9.

Dot plots showing data distribution in the cohort (n = 80) for the acetaminophen metabolites, four primary bile acids, and GSH. The red dots show the metabolite level for either volunteer 3976 or volunteer 3958. The open dots show the data distribution for the rest of the cohort. An explanation of the plots is provided in the legend for Fig. 3. The first and second numbers within the parenthesis are the z-score and P value, respectively.

Fig. 10.

Fig. 10.

Relative serum level of atorvastatin and daily dose among the 10 volunteers taking Lipitor.

Fig. S1.

Fig. S1.

Data distribution for gut bacteria-derived metabolites in the cohort. The red dots show the metabolite level for volunteer 3930. The open dots show the data distribution for the rest of the cohort.

Similar articles

Cited by

References

    1. Yang Y, et al. Clinical whole-exome sequencing for the diagnosis of mendelian disorders. N Engl J Med. 2013;369(16):1502–1511. - PMC - PubMed
    1. Berg JS, et al. Next generation massively parallel sequencing of targeted exomes to identify genetic mutations in primary ciliary dyskinesia: Implications for application to clinical testing. Genet Med. 2011;13(3):218–229. - PMC - PubMed
    1. Maxmen A. Exome sequencing deciphers rare diseases. Cell. 2011;144(5):635–637. - PubMed
    1. Ng SB, et al. Exome sequencing identifies the cause of a mendelian disorder. Nat Genet. 2010;42(1):30–35. - PMC - PubMed
    1. Roach JC, et al. Analysis of genetic inheritance in a family quartet by whole-genome sequencing. Science. 2010;328(5978):636–639. - PMC - PubMed

MeSH terms

LinkOut - more resources