Detecting haplotype effects in genomewide association studies

B. Huang,C. Amos,Danyu Lin
DOI: https://doi.org/10.1002/GEPI.20242
2007-12-01
Genetic Epidemiology
Abstract:The analysis of genomewide association studies requires methods that are both computationally feasible and statistically powerful. Given the large‐scale collection of single nucleotide polymorphisms (SNPs), it is desirable to explore the information contained in their interrelationships. In particular, utilizing haplotypes rather than individual SNPs and accounting for correlations of polymorphisms in adjustment for multiple testing can lead to increased power. We present a statistically powerful and numerically efficient method based on sliding windows of adjacent SNPs to detect haplotype‐disease association in genomewide studies. This method consists of an efficient algorithm to calculate a proper likelihood‐ratio statistic for any given window of SNPs, along with an accurate and efficient Monte Carlo procedure to adjust for multiple testing. Simulation studies using the HapMap data showed that the proposed method performs well in realistic situations. We applied the new method to a case‐control study on rheumatoid arthritis and identified several loci worthy of further investigations. Genet. Epidemiol. 2007. © 2007 Wiley‐Liss, Inc.
What problem does this paper attempt to address?