| tscc {gap} | R Documentation |
Power calculation for two-stage case-control design
Description
Power calculation for two-stage case-control design
Usage
tscc(model, GRR, p1, n1, n2, M, alpha.genome, pi.samples, pi.markers, K)
Arguments
model |
any in c("multiplicative","additive","dominant","recessive"). |
GRR |
genotype relative risk. |
p1 |
the estimated risk allele frequency in cases. |
n1 |
total number of cases. |
n2 |
total number of controls. |
M |
total number of markers. |
alpha.genome |
false positive rate at genome level. |
pi.samples |
sample% to be genotyped at stage 1. |
pi.markers |
markers% to be selected (also used as the false positive rate at stage 1). |
K |
the population prevalence. |
Details
This function gives power estimates for two-stage case-control design for genetic association.
The false positive rates are calculated as follows,
P(|z1|>C1)P(|z2|>C2,sign(z1)=sign(z2))
and
P(|z1|>C1)P(|zj|>Cj||z1|>C1)
for replication-based and joint analyses, respectively; where C1, C2, and Cj are threshoulds at stages 1, 2 replication and joint analysis,
z1 = z(p1,p2,n1,n2,pi.samples)
z2 = z(p1,p2,n1,n2,1-pi.samples)
zj = sqrt(pi.samples)*z1+sqrt(1-pi.samples)*z2
Value
The returned value is a list containing a copy of the input plus output as follows,
model any in c("multiplicative","additive","dominant","recessive").
GRR genotype relative risk.
p1 the estimated risk allele frequency in cases.
pprime expected risk allele frequency in cases.
p expected risk allele frequency in controls.
n1 total number of cases.
n2 total number of controls.
M total number of markers.
alpha.genome false positive rate at genome level.
pi.samples sample% to be genotyped at stage 1.
pi.markers markers% to be selected (also used as the false positive rate at stage 1).
K the population prevalence.
C threshoulds for no stage, stage 1, stage 2, joint analysis.
power power corresponding to C.
Note
solve.skol is adapted from CaTS.
Author(s)
Jing Hua Zhao
References
Skol AD, Scott LJ, Abecasis GR, Boehnke M (2006). “Joint analysis is more efficient than replication-based analysis for two-stage genome-wide association studies.” Nat Genet, 38(2), 209-13. ISSN 1061-4036 (Print) 1061-4036, doi:10.1038/ng1706.
Examples
## Not run:
K <- 0.1
p1 <- 0.4
n1 <- 1000
n2 <- 1000
M <- 300000
alpha.genome <- 0.05
GRR <- 1.4
p1 <- 0.4
pi.samples <- 0.2
pi.markers <- 0.1
options(echo=FALSE)
cat("sample%,marker%,GRR,(thresholds x 4)(power estimates x 4)","\n")
for(GRR in c(1.3,1.35,1.40))
{
cat("\n")
for(pi.samples in c(1.0,0.5,0.4,0.3,0.2))
{
if(pi.samples==1.0) s <- 1.0
else s <- c(0.1,0.05,0.01)
for(pi.markers in s)
{
x <- tscc("multiplicative",GRR,p1,n1,n2,M,alpha.genome,
pi.samples,pi.markers,K)
l <- c(pi.samples,pi.markers,GRR,x$C,x$power)
l <- sprintf("%.2f %.2f %.2f, %.2f %.2f %.2f %.2f, %.2f %.2f %.2f %.2f",
l[1],l[2],l[3],l[4],l[5],l[6],l[7],l[8],l[9],l[10],l[11])
cat(l,"\n")
}
cat("\n")
}
}
options(echo=TRUE)
## End(Not run)