Table 2 

Impact of replicate arrays on effective sample size (N*) and minimum detectable odds ratio (MDOR) in poolingGWAS. 

Arrays per pool 
Case pool (RSS, N*) 
Control pool (RSS, N*) 
MDOR at 80% (p = 0.29) 
MDOR at 80% (p = 0.10) 


24 
0.95, 284 
0.84, 837 
1.33 
1.51 
12 
0.90, 269 
0.72, 720 
1.35 
1.54 
6 
0.81, 244 
0.56, 562 
1.38 
1.58 
3 
0.69, 206 
0.39, 391 
1.44 
1.70 
Individual Genotyping 
1, 300 
1, 1000 
1.32 
1.49 


This table compares the minimum detectable odds ratios (MDOR) at 80% power for a theoretical pooling experiment with 300 cases and 1000 controls, given a DNApooling strategy where 24, 12, 6, or 3 Illumina 660Quad replicate arrays are used to allelotype each DNA pool (case and control). The equivalent individual genotyping experiment is given for reference. Relative sample size (RSS) and effective sample size (N*) are generated by PoolingPlanner assuming var(e_{array})= 3.3 × 10^{4}, var(e_{construction})= 9.9 × 10^{5}, and an average minor allele frequency of 0.29. MDOR at 80% power were calculated using Quanto [21] assuming an unmatched casecontrol design testing for geneonly effects using a logadditive model, where the incidence of the case phenotype is 0.02% and the risk allele, p, is set to 0.29 or 0.10. 

Earp et al. BMC Medical Genomics 2011 4:81 doi:10.1186/17558794481 