Predictive and Aetiological Potential of Polygenic Strata with Extreme and Moderate Disease Risks Predictions: Case Study of Coeliac Disease GWAS

We show by using example of coeliac disease (CD) that a genomic risk assessment could significantly improve efficiency of disease diagnosing. It can detect novel highly deleterious rare variants (penetrance 100%, frequency [~]1:6,700) as well as common protective variants (penetrance 0.03%, frequency [~]1:3). However, the major translational gains with potential for multi-billion-dollar cost savings in Australia or USA alone, could be in assessing patients in cohorts with moderately elevated CD risk (3% -10%) exhibiting clinical symptoms or with family history of CD. The gains result from judicious re-direction of expensive confirmatory testing towards [~]30% of the cohort with the highest likelihood of the condition ([~]90% of cohort CD cases), while avoiding costs, inconvenience and risk of side-complications for the remaining majority of [~]70%.\n\nWe build our estimates using concrete results of CD Genome Wide Association Studies (GWAS) already in the public domain1-4. The largest of five Genomic Risk Score (GRS) models1 considered here deploys 228 directly genotyped Simple Nucleotide Polymorphisms (SNPs), while the simplest2 uses only 6 SNPs. Thus, a DNA profile supporting all these models can be easily accommodated on any commodity, Direct-to-Consumer5 (DTC), saliva-based genotyping platform. Once generated, such a generic profile of over 600,000 SNPs could assist medical practitioners in diagnosing this as well as thousands of other diseases on demand, virtually genotyping cost free.