SKAT.matrixfree {bigQF}R Documentation

Make 'matrix-free' object for SKAT test

Description

'Matrix-free' or 'implicit' linear algebra uses a matrix only through the linear operations of multiplying by the matrix or its transpose. It's suitable for sparse matrices, and also for structured matrices that are not sparse but still have fast algorithms for multiplication. The Sequence Kernel Association Test is typically performed on sparse genotype data, but the matrix involved in computations has been centered and is no longer sparse.

Usage

SKAT.matrixfree(G,weights=function(maf) dbeta(maf,1,25), model=NULL,...)
## S3 method for class 'lm'
SKAT.matrixfree(G,weights=function(maf) dbeta(maf,1,25), model=NULL,...)
## S3 method for class 'glm'
SKAT.matrixfree(G,weights=function(maf) dbeta(maf,1,25), model=NULL,...)
## S3 method for class 'lmekin'
SKAT.matrixfree(G,weights=function(maf) dbeta(maf,1,25), model=NULL, kinship,...)

Arguments

G

A 0/1/2 matrix whose columns are markers and whose rows are samples. Should be mostly zero.

weights

A weight function used in SKAT: the default is the standard one.

model

A linear model, generalised linear model, or lmekin object used for adjustment, or NULL

kinship

A (sparse) kinship matrix. The model will have used 2*kinship in its varlist argument.

...

to keep CMD check happy

Details

If the adjustment model is NULL the object contains the trace of the underlying quadratic form, if the adjustment model is not NULL the trace will be estimated using Hutchinson's randomised estimator inside pQF. The lmekin method calls famSKAT.

Value

An object of class matrixfree

Author(s)

Thomas Lumley

References

Lee, S., with contributions from Larisa Miropolsky, and Wu, M. (2015). SKAT: SNP-Set (Sequence) Kernel Association Test. R package version 1.1.2.

Lee, S., Wu, M. C., Cai, T., Li, Y., Boehnke, M., and Lin, X. (2011). Rare-variant association testing for sequencing data with the sequence kernel association test. American Journal of Human Genetics, 89:82-93.

Wu, M. C., Kraft, P., Epstein, M. P., Taylor, D. M., Channock, S. J., Hunter, D. J., and Lin, X. (2010). Powerful SNP set analysis for case-control genome-wide association studies. American Journal of Human Genetics, 86:929-942.

See Also

pQF sparse.matrixfree

Examples

data(sequence)
skat<-SKAT.matrixfree(sequence)
skat$trace
pQF(c(33782471,7e7,1e8), skat,n=100, tr2.sample.size=500)

data(SKAT.example)

skat1mf <- SKAT.matrixfree(SKAT.example$Z)
(Q<-skat1mf$Q(SKAT.example$y.c))
all.equal(as.numeric(Q), 234803.786)

## correct value is 0.01874576
pQF(Q, skat1mf, neig=4, convolution.method="integration")


skat3mf <- SKAT.matrixfree(SKAT.example$Z, model=lm(y.c~X, data=SKAT.example))
(Q<-skat3mf$Q())
all.equal(Q, 298041.542)

## correct value is 0.002877041
pQF(Q, skat3mf, neig=4, convolution.method="integration")



[Package bigQF version 1.6 Index]