GWAS in the Quebec Founder Population
Genizon BioSciences (www.genizon.com) faces the challenges of GWAS in unique ways. All samples collected come from the French Canadian population of Quebec, known as the Quebec Founder Population (QFP), which is genetically less heterogeneous than general populations. Thus, fewer resources are required for well-powered GWAS. In Genizon’s experience, when using the QFP, only 500 trios or 750 case-controls provide good power for less complex diseases, such as Crohn’s disease (Figure 2) and psoriasis. On the other hand, 1,000–1,500 cases and controls are required for more complex diseases, like type II diabetes and schizophrenia. For general populations, it is usually believed that at least 2,000–5,000 cases and controls may be required.
Use of genetic data from the QFP has also enabled the development of a comprehensive genome-wide SNP map, representative of linkage disequilibrium in this population. Genizon has customized an Illumina HumanHap300 BeadChip (317,000 SNPs) by adding 57,000 SNPs distributed according to the specific genetic profile of this population. Using the Illumina BeadStation genotyping platform, the company routinely achieves call rates and accuracy rates that exceed 99.9%.
To enhance association signals and increase the distance at which they can be detected, Genizon has developed and implemented efficient multimarker haplotype analyses that permit their implementation at the genome-scan stage. Haplotype patterns are estimated using a modified EM algorithm with high accuracy using family trio or case-control data. A high-speed algorithm LDSTATS is then used to rapidly compare the frequency of single markers and haplotypes between patients and controls across the entire genome. The use of haplotype analyses can result in more significant association signals and was found to be critical to the discovery of many of the signals identified in Genizon’s studies.
Of primary importance, due to the large number of markers tested, Genizon assesses the genome-wide significance of the study by permutation analyses. In addition, methods are used to identify gene-gene interactions, including conditional analyses and a multifactorial approach, Random Forests. Subphenotypes and gender-specific analyses are routinely used to provide more homogeneous subsets of patients thereby increasing effect size and statistical significance. These additional analyses have resulted in multiple additional discoveries of important disease-susceptibility genes.
In Genizon’s experience, the use of the QFP, derived from a small group of founders 10–20 generations ago that has expanded in relative isolation to a large population today, coupled with advanced genetic analysis algorithms has had a dramatic impact on the power to discover disease genes.
To date, Genizon has successfully completed eight GWAS, including Crohn’s disease, psoriasis, asthma, endometriosis, ADHD, schizophrenia, longevity, and male pattern baldness. These studies have resulted in the identification of disease-susceptibility genes, their associated pathways, and potential new therapeutic targets.