twoStageGwasPower: thresholds and power for two-stage gwas

In twoStageGwasPower: Compute thresholds and power for two-stage gwas

Description

Computes the critical values for a two-stage gwas association study and compute power for replication and joint analyses.

Usage

twoStageGwasPower(pD, pG, grr, inheritance = "multiplicative", pi.samples, pi.markers, alpha.marker, n.cases, n.controls)

Arguments

pD	Probability of disease in the population (prevalence)
pG	Frequency of disease allele in the population
grr	Genotypic relative risk
inheritance	Inheritance type ("dominant", "recessive", "multiplicative", or "additive"
pi.samples	Proportion of samples genotyped in Stage 1
pi.markers	Proportion of markers genotyped in Stage 2
alpha.marker	Significance level used for each marker, accounting for multiple comparisons among a large number of markers
n.cases	Number of cases
n.controls	Number of controls

Details

This function computes the critical values and powers of the replication and joint methods of analyzing a two-stage GWAS design. Details may be found in Skol AD, Scott, LJ, Abecasis GR, Boehnke M (2006)

Value

A list containing:

power.SingleStage	Power of a one stage design
power.joint	Power of a joint analysis
power.rep	Power of a replication analysis (based only on the second stage markers)
c1	Stage one threshold
c2	Replication (stage two) threshold
c.joint	Joint analysis threshold
c.singleStage	Single stage design threshold
penetrance.GG	Penetrance of the GG genotype (homozygous for disease allele)
penetrance.Gg	Penetrance of the Gg genotype
penetrance.gg	Penetrance of the gg genotype
p0	disease allele frequency in controls
p1	disease allele frequency in cases
p.stageOne	probability that associated markers will be followed up in Stage 2
savings	reduction in genotyping using two-stage design as compared to the single-stage design

Author(s)

Dirk F Moore <[e-mail]>

References

Skol AD, Scott, LJ, Abecasis GR, Boehnke M (2006) Nature Genetics doi:10.1038/ng1706

http://www.sph.umich.edu/csg/abecasis/CaTS/

Examples

# prevalence of disease is 0.10, the allele frequency is 0.40,
#  a multiplicative model with 0.40 samples in the first stage and
#  10% of the markers selected for Stage 2. There are 1000 cases
#  and 1000 controls, 300,000 markers, with a genome-wide alpha of 0.05

power.gwas.out <- twoStageGwasPower(pD=0.10, pG=0.40, grr=1.40, 
                 inheritance="multiplicative", pi.samples=0.40, pi.markers=0.10, 
                 alpha.marker=0.05/300000, n.cases=1000, n.controls=1000)
power.gwas.out

# Same, but with 1% of markers selected for Stage 2
power.gwas.out2 <- twoStageGwasPower(pD=0.10, pG=0.40, grr=1.40, 
             inheritance="multiplicative", pi.samples=0.40, pi.markers=0.010, 
             alpha.marker=0.05/300000, n.cases=1000, n.controls=1000)
power.gwas.out2

# Same, but a dominant model with 4000 controls and 2000 cases
power.gwas.out3 <- twoStageGwasPower(pD=0.10, pG=0.40, grr=1.40, 
           inheritance="dominant", pi.samples=0.40, pi.markers=0.10, 
           alpha.marker=0.05/300000, n.cases=2000, n.controls=4000)
power.gwas.out3

Example output

Power for one-stage design 0.7437531 
Power for replication analysis 0.4561015 
Power for joint analysis 0.7419231 

Threshold for one-stage design 5.233126 
Threshold for stage one 1.644854 
Threshold for replication design 4.649133 
Threshold for joint design 5.230999 

Penetrance for GG 0.1456599 
Penetrance for Gg 0.1040428 
Penetrance for gg 0.07431629 

Disease allele frequencies in cases 0.4827586 
Disease allele frequencies in controls 0.3908046 
Probability that associated markers will be followed up in stage two 0.9811959 

Reduction in genotyping from two-stage design 0.54 
Power for one-stage design 0.7437531 
Power for replication analysis 0.5768826 
Power for joint analysis 0.7110829 

Threshold for one-stage design 5.233126 
Threshold for stage one 2.575829 
Threshold for replication design 4.14941 
Threshold for joint design 5.204231 

Penetrance for GG 0.1456599 
Penetrance for Gg 0.1040428 
Penetrance for gg 0.07431629 

Disease allele frequencies in cases 0.4827586 
Disease allele frequencies in controls 0.3908046 
Probability that associated markers will be followed up in stage two 0.8745411 

Reduction in genotyping from two-stage design 0.594 
Power for one-stage design 0.5358034 
Power for replication analysis 0.2867125 
Power for joint analysis 0.5341269 

Threshold for one-stage design 5.233126 
Threshold for stage one 1.644854 
Threshold for replication design 4.649133 
Threshold for joint design 5.230999 

Penetrance for GG 0.111465 
Penetrance for Gg 0.111465 
Penetrance for gg 0.07961783 

Disease allele frequencies in cases 0.4458599 
Disease allele frequencies in controls 0.3949045 
Probability that associated markers will be followed up in stage two 0.9574386 

Reduction in genotyping from two-stage design 0.54 

twoStageGwasPower documentation built on May 29, 2017, 12:11 p.m.