Statistical Power for Genome-Wide Association Analysis of Case-Control Studies on Rare Serious Adverse Drug Reactions

• Activity Number: 515
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Wencan Zhang*+ and Tatsuya Ando and Robert Dillard and Leonardo Sahelijo and Shyh-yuh Liou and Hiroyoshi Toyoshiba
• Companies: Takeda Global Research & Development Center, Inc. and Takeda Pharmaceutical Company Limited and Takeda Global Research & Development Center, Inc. and Takeda Global Research & Development Center, Inc. and Takeda Pharmaceutical Company Limited and Takeda Pharmaceutical Company Limited
• Keywords: Rare Serious Adverse Drug Reactions ; GWAS ; Univariate Statistical Power ; Sequential Genome-Wide Multivariate Analysis ; Limited Case Sample Size ; Public Control Data Sets

Abstract

Understanding the genome-wide association of the occurrence of serious adverse drug reactions (SADRs) for a drug is a significant challenge in drug development and drug therapy. Collection of blood samples for such testing "post facto" is inefficient and costly. Consequently, limiting collection of "post facto" samples only from subjects with SADRs in conjunction with publicly available SNP control data sets requires careful power assessment. Power under conventional univariate methods(Menashe et al 2008) for different levels of multiple factors was assessed, assuming the prevalence of a SADR of 0.01, 1 million SNPs, recessive allele frequency of 0.1 or 0.2 for both gene and marker, and a ratio of 3 to 30 on relative risk (RR) for associated homozygous alleles SNP genotypes. Power from a computer simulated sequential genome-wide multivariate analysis (Kelly et al 2006) is also presented.