Genetics of human metabolism: an update

Abstract

Genome-wide association studies with metabolomics (mGWAS) identify genetically influenced metabotypes (GIMs), their ensemble defining the heritable part of every human's metabolic individuality. Knowledge of genetic variation in metabolism has many applications of biomedical and pharmaceutical interests, including the functional understanding of genetic associations with clinical end points, design of strategies to correct dysregulations in metabolic disorders and the identification of genetic effect modifiers of metabolic disease biomarkers. Furthermore, it has been shown that GIMs provide testable hypotheses for functional genomics and metabolomics and for the identification of novel gene functions and metabolite identities. mGWAS with growing sample sizes and increasingly complex metabolic trait panels are being conducted, allowing for more comprehensive and systems-based downstream analyses. The generated large datasets of genetic associations can now be mined by the biomedical research community and provide valuable resources for hypothesis-driven studies. In this review, we provide a brief summary of the key aspects of mGWAS, followed by an update of recently published mGWAS. We then discuss new approaches of integrating and exploring mGWAS results and finish by presenting selected applications of GIMs in recent studies.

Figure 1.

Manhattan plot of the mGWAS by Shin et al. (23). Upward pointing P-values: TwinsUK cohort, downward pointing P-values: KORA population study. Only SNPs with association to raw metabolites (P < 10⁻⁶) are displayed (no ratios). The green line indicates the genome-wide significance cutoff for the P-value (P < 10⁻¹⁰). Loci that reach genome-wide significance in either cohort are indicated by a short vertical black line. Loci with P-values <10⁻³⁰ are indicated with a red symbol. Loci that are further discussed in Figure 3 are highlighted and annotated [figure adapted from the Supplementary Material, Fig. S2 by Shin et al., Nature Genetics, 2014 (23)].

Figure 2.

Network integrating gene–metabolite associations and metabolite–metabolite correlations. Individual metabolites are lumped by pathway (colored circles) and colored by their general metabolic properties (see legend). Genetic loci (gray diamonds) are annotated by the gene that is most likely affected by the variant. Green edges between loci and metabolites represent significant genetic associations with metabolic traits. Gray edges between metabolites represent significant partial correlations between metabolic traits. The highlighted sub-network (shaded box) is further discussed in Figure 3. The full network is freely accessible in digital format at http://gwas.eu/si [figure adapted from Figure 2 by Shin et al., Nature Genetics, 2014 (23)].

Figure 3.

Example of the integration of mGWAS results in a biomedical context using data from different sources. This figure displays a ‘Cardiovascular disease and hypertension metabolic sub-network’, annotated based on correlations between molecular relationships and expert knowledge on blood pressure regulation, blood coagulation and known molecular risk factors for cardiovascular disease and hypertension. Metabolites (circles) and genes (diamonds) of the fibrinogen cleavage (left) and the kininogen/kinin system (right) and their interconnections were derived from the Shin et al. (see shaded box in Fig. 2) data. Gray nodes and edges display annotations of biochemical function based on expert knowledge (47). Colored nodes and edges correspond to reported associations based on genome-wide studies for blood pressure regulation (orange), blood coagulation (blue) and cholesterol levels (purple). This figure was first published as the Supplementary Material, Fig. S5 by Shin et al., Nature Genetics, 2014 (23).