FindOptPower: search for the optimal pooled design

In hiPOD: hierarchical Pooled Optimal Design

Description

Perform a grid search over potential design space, and find the predicted power and validity of the designs.

Usage

FindOptPower(cost, sample.size, MAF, OR, error, costPerExp = 18915, costPerPool = 970, costPerX = 300, lower.P = 20, upper.P = 400, lower.N.p = 2, upper.N.p = 200, lower.Xmean = 4, upper.Xmean = 1280, sig.level = 0.05, Number.Grids = 100)

Arguments

cost	cost constraint of the study
sample.size	sample size constraint of the study
MAF	assumed MAF of the variant of interest
OR	assume effect size (odds ratio) of the variant of interest
error	assume sequencing error rate
costPerExp	cost per experiment
costPerPool	cost per pool
costPerX	cost per 1X coverage
lower.P	lower bound of number of pools in potential consideration
upper.P	upper bound of number of pools in potential consideration
lower.N.p	lower bound of number of pool size in potential consideration
upper.N.p	upper bound of number of pool size in potential consideration
lower.Xmean	lower bound of number of coverage per pool in potential consideration
upper.Xmean	upper bound of number of coverage per pool in potential consideration
sig.level	significance level of the statistic test, usually 0.05 for a single test
Number.Grids	number of grids in the search space, preset as 100

Details

Given the research question and the study constraints, this function calculates the power and validity of all the potential pooled designs.

Value

Returns a list:

cost
sample.size
constraint.set
scenario.set
designs	the potential designs, validity and power

Author(s)

Wei E. Liang

See Also

PlotOptPower, ShowOptDesign

Examples

# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
######## Example 1: A simple example, with very rough grid points (only 20X20 grid points)
# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #

##### Find the optimal design
example.1 <- FindOptPower(cost=452915, sample.size=5000, MAF=0.03, OR=2, error=0.01, upper.P=200, Number.Grids=20)

##### assign a directory to store the contour plots
##### with your own choice
proj.Dir <- paste(getwd(), "/hiPOD_examples", sep="")
if(!file.exists(proj.Dir)) dir.create(proj.Dir)

##### Inferences on the optimal designs
PlotOptPower(example.1, save.contour=FALSE, plot.3d=FALSE)
# # snapshot3d(filename = paste(proj.Dir, "/example1_3d.bmp", sep=""))
ShowOptDesign(example.1, 5)

hiPOD documentation built on May 1, 2019, 6:34 p.m.