power {pbatR}R Documentation

Power And Sample Size Calculations

Description

Does power and sample size calculations. Note that this may be very highly pbat version specific, and more finicky than other methods in this package, as it requires that the menuing interface does not change.

The Windows version requires a slight hack to control PBAT, which appears to be stable, but may have a race condition on heavily loaded systems. Unix/Linux users should not encounter this.

'pbat.power()' starts the GUI interface (strongly recommended) 'pbat.binaryFamily(...)' is for family based designs with binary 'pbat.continuous(...)' is for family based designs with continuous 'pbat.caseControl(...)' is for case/control designs 'pbat.popQuant(...)' is for population designs with quantitative

Currently less extensive range checking is done in these routines unlike the rest of the routines in pbatR.

Usage

pbat.power()

pbat.binaryFamily(
           numOffspring=1, missingParents=0, numFam=0,
           addiOffspringPheno=1, ## only when you have missing parents
           ascertainment="unaffected",
           model="additive", model.afreq=0.2, model.incrAfreq=0,
           model.disLocIsMarker=TRUE,
           
           model.popPrev=NULL,  ## Options 1, 3, & 4
           model.genAF=NULL,    ## Option 1
           model.penAA=NULL, model.penAB=NULL, model.penBB=NULL, ## Option 2
           model.OR=NULL,       ## Option 3
           model.aOR=NULL,      ## Option 4
           
           stat.sigLevel=0.01,
           stat.offset="",  ## defaults to population prevalence
           comp="numerical",
           log="pbatLog.txt")

pbat.continuousFamily(
           numOffspring=1, missingParents=0, numFam=0,
           addiOffspringPheno, ## only when you have missing parents
           ascertainment="unaffected",
           model="additive", model.afreq=0.1, model.incrAfreq=0,
           model.disLocIsMarker=FALSE,                      
           model.heritability=0.1, model.afreqMarker=0,
           model.prDiseaseGivenMarker=1,
           stat.sigLevel=0.05, stat.offset="",
           comp="numerical",
           log="pbatLog.txt")

pbat.caseControl(
           model="additive", model.minafreq=0.1, model.incrAfreq=0.1,
           model.prevalence=0.1,
           model.ORofABvsBB=NULL, # Option 1 - default 1.5
           model.aOR=NULL,      # Option 2 - default 1.481
           comp.cases=500, comp.controls=500, comp.caseControlRatio=1.5,
           comp.power=0.8, comp.sigLevel=0.05, comp.numSim=1000,
           mode="power",
           log="pbatLog.txt")

pbat.popQuant(
           model="additive",
           model.minafreq=0.2, model.incrAfreq=0.1,
           model.heritability=0.001,
           comp.numProbands=2000,
           comp.power=0.8, comp.sigLevel=0.05, comp.numSim=10000,
           mode="power",
           log="pbatLog.txt")

Arguments

numOffspring Family - number of offspring
missingParents Family - number of missing parents (0,1,2)
numFam Family - number of families
addiOffspringPheno Only used when you have missing parents; additional offspring phenotypes. 1 for yes, 0 for no.
ascertainment 'unaffected', 'affected', or 'not applicable'
model 'additive', 'dominant', 'recessive' or (only for binary / case control) 'multi'
model.afreq allele frequency
model.incrAfreq increment allele frequency
model.disLocIsMarker TRUE/FALSE - whether the disease locus is the same as the marker locus
model.popPrev population prevalence
model.prevalence population prevalence
model.genAF genetic attributable fraction of the gene
model.penAA penetrance of AA genotype
model.penAB penetrance of AB genotype
model.penBB penetrance of BB genotype
model.OR odds ratio
model.aOR allelic odds ratio
model.heritability heritibility
model.afreqMarker marker allele frequency
model.prDiseaseGivenMarker Pr(Disease|marker)
model.ORofABvsBB odds ratio of AB to BB
model.minafreq minimum allele frequency
stat.sigLevel significance level
stat.offset offset; defaults to population prevalence
comp 'numerical':numerical integration, 'approximation':approximation, 'simulation':simulation
comp.cases number of cases
comp.controls number of controls
comp.caseControlRatio case control ratio
comp.power power
comp.sigLevel significance level
comp.numSim number of simulations to run
comp.numProbands number of probands
mode 'power' for power result given parameter specification, 'ss' for sample size result given parameter specification; note that not all parameters will be used for 'power' and 'ss'.
log logfile to write to (results will be stored here) - this will be over-written if the file exists

References

http://www.biostat.harvard.edu/~clange/default.htm

http://www.people.fas.harvard.edu/~tjhoffm/pbatR.html

See Also

pbat, pbat.last


[Package pbatR version 1.1 Index]