A multi-institutional team of researchers, led by Basic Sciences faculty member Bingshan Li (Molecular Physiology & Biophysics), has developed a new framework that can help researchers learn more from genome-wide association studies (GWAS) than previously was possible. Their work was published in Nature Neuroscience.
The purpose of GWAS is to scan the genomes of many individuals to find genetic variations associated with a particular disease. In the case of schizophrenia, GWAS have identified over 100 disease-associated loci, but limitations of the technique – namely, that GWAS loci usually cover multiple candidate genes and that the gene that constitutes a disease risk might actually be megabases away from an identified locus – means that it has been exceedingly difficult to ascribe biological meaning to the data.
To overcome these limitations, Li and colleagues designed integrative risk gene selector (iRIGS), a Bayesian framework that probabilistically infers risk genes driving GWAS signals by considering whether a particular gene has been identified as a risk gene through multi-omics approaches (data derived from multiple -omics methods, like transcriptomics and metabolomics) as well as the relationships of genes in biological networks.
As proof of principle, the investigators applied their methodology to a GWAS schizophrenia dataset. They found that the iRIGS-predicted risk genes are highly consistent with the leading pathophysiological hypotheses of schizophrenia and that they help explain the high disease heritability. Additionally, iRIGS identified a number of risk genes that are already targeted by FDA-approved, clinically investigational, or preclinical drugs.
GWAS has allowed scientists to generate large datasets from multiple individuals, but thanks to its integration with other data sources (e.g. -omics), iRIGS has the potential to explain and extrapolate on some of these data, helping researchers interpret and act on obtained results.