Chapter 4 Learning the Optimal in GWAS through hierarchical SNP aggregation

The present chapter proposes a block-wise approach for GWAS analysis which leverages the LD structure among the genomic variants to reduce the number of hypothesis testing. We named this method LEOS for LEarning the Optimal Scale in GWAS. Section 4.1 introduces some related works that have been studied to develop our methodology. The method is presented in Section 4.2. In Section 4.3, we compare our method in different scenarios with the baseline approach, i.e. univariate hypothesis testing (Purcell et al. 2007) and with the logistic kernel machine method presented in Section 3.7.2 on both synthetic and real datasets from the Wellcome Trust Case Control Consortium (WTCCC 2007) and on ankylosing spondylitis data (International Genetics of Ankylosing Spondylitis Consortium (IGAS) et al. 2013). Finally, an example of an application using the generalized additive models in the context of GWAS is exposed in Section 4.5.

References

International Genetics of Ankylosing Spondylitis Consortium (IGAS), Adrian Cortes, Johanna Hadler, Jenny P. Pointon, Philip C. Robinson, and others. 2013. “Identification of Multiple Risk Variants for Ankylosing Spondylitis Through High-Density Genotyping of Immune-Related Loci.” Nature Genetics 45 (7): 730–38.

WTCCC. 2007. “Genome-Wide Association Study of 14,000 Cases of Seven Common Diseases and 3,000 Shared Controls.” Nature 447 (7145): 661–78.