With a small vial of blood, researchers showed in 2009 that they could determine a healthy person's metabolic phenotype. That is, by coupling microarray readouts to mass spectrometry data derived from the same sample, they were able to identify variants in metabolism-related genes that conferred certain metabotypes.
As it has become increasingly clear that complex diseases are not caused by genetic variation alone, Karsten Suhre and his colleagues have opted to examine metabolites, as they are essentially the molecular middlemen between genotype and phenotype. In a 2009 Nature Genetics paper, Suhre — who was then at Germany's Ludwig Maximilians-Universität — and his colleagues reported more than a dozen genetic loci that are significantly associated with metabolite concentration ratios in human blood.
"In metabolomics, you are looking at things which are intermediates," says Suhre, who now directs the bioinformatics core at Weill Cornell Medical College in Qatar. By coupling genome-wide association studies to metabolomics, "the key is we are looking at intermediate phenotypes," he says, adding that "genes tell you what your body can do, but your metabolism tells you what really happens there."
When combined, Suhre says GWAS and metabolomics provide researchers with "a much bigger picture of the patient." Using both approaches is "all about visibility. ... We now have the possibility to see a lot more things in the patient — things that [could] link genetic variants and complex disease," he adds.
To investigate the detoxification capacity of the human body, the team applied its GWAS for metabolomics approach to urine. Because of the nature of urine, the researchers used NMR spectroscopy in place of mass spectrometry to examine the samples. In May, Suhre and his colleagues reported in Nature Genetics their identification of five genetic variants at loci associated with the concentration of specific metabolites present in human urine. One is a SNP in the coding sequencing of the alanine-glyoxylate aminotransferase-2 gene, or AGXT2, which the team identified as the genetic basis of hyper-β-aminoisobutyric aciduria — a condition marked by elevated levels of of β-aminoisobutyratic acid in the urine.
First recognized in 1951, hyper-β-aminoisobutyric aciduria is one of the most common monogenetic phenotypes in humans. It was not until Suhre and his colleagues connected the metabolite to a genotype, though, that the cause of this Mendelian condition was resolved. "For 50 years ... nobody knew what gene it was," he says. "We weren't looking for it; it just popped out."
By taking a hypothesis-free approach — simply screening healthy population study participants for the presence of millions of SNPs and hundreds of metabolites — Suhre says his team has been happily surprised by the amount of overlap they've observed between genotype and the metabolic traits of blood and urine.
"We just measure everything we can measure. In the end, you do the statistical associations and find some SNP goes with some metabolite, so that's nice. But then you put the name on it and you find out that the gene is actually the enzyme which acts on the metabolite you measured," he says.
So far, Suhre's team has identified several "genes where there is a perfect match between the function of the gene and the metabolite we -measure," he says. Aside from the SNP in AGXT2, Suhre's team reported four additional variants that are associated with concentrations of certain metabolites in human urine. That these functional connections have all simply fallen into place is "for me, -really most amazing," he adds.
Suhre and his colleagues are now applying their approach to an expanded panel of metabolites to see what other associations may fall out. "With this technique, we hope to find even more of these genetic variants," he says.