R: SKAT for the combined effect of common and rare variants

SKAT_CommonRare {SKAT}

R Documentation

SKAT for the combined effect of common and rare variants

Description

Sequence Kernel association test for the combined effect of common and rare variants.

Usage


	SKAT_CommonRare(Z, obj, weights.beta.rare=c(1,25)
	, weights.beta.common=c(0.5,0.5), weights=NULL
	, method="C", r.corr.rare=0, r.corr.common=0, CommonRare_Cutoff=NULL
	, test.type="Joint", is_dosage=FALSE, missing_cutoff=0.15
	, estimate_MAF=1, SetID1=NULL)


	SKAT_CommonRare.SSD.OneSet(SSD.INFO
	, SetID, obj, ..., obj.SNPWeight=NULL)

	SKAT_CommonRare.SSD.OneSet_SetIndex(SSD.INFO
	, SetIndex, obj, ..., obj.SNPWeight=NULL )

Arguments

`Z`	a numeric genotype matrix with each row as a different individual and each column as a separate gene/snp. Each genotype should be coded as 0, 1, 2, and 9 (or NA) for AA, Aa, aa, and missing, where A is a major allele and a is a minor allele. Missing genotypes will be imputed by the simple Hardy-Weinberg equilibrium (HWE) based imputation.
`obj`	an output object of the SKAT_Null_Model function.
`weights.beta.rare`	a numeric vector of parameters of beta weights for rare variants (default=c(1,25)).
`weights.beta.common`	a numeric vector of parameters of beta weights for common variants (default=c(0.5,0.5)).
`weights`	a numeric vector of weights for both common and rare variants. When it is NULL, the beta weight with the “weights.beta.rare” and “weights.beta.common” parameter are used. When method ="C", the coefficient to combine common and rare variants test statistics will be calculated after applying the weights to variants.
`method`	a method to combine common and rare variant effects (default="C"). "C" represents the combined sum test, and "A" represents the adaptive sum test. "AR" represents a different type of adaptive test in which common variants are projected over rare variants.
`r.corr.rare`	the `\rho` parameter for rare variants (default= 0). `\rho` =0 and 1 indicate SKAT and Burden test, respectively
`r.corr.common`	the `\rho` parameter for common variants (default= 0). `\rho` =0 and 1 indicate SKAT and Burden test, respectively
`CommonRare_Cutoff`	MAF cutoff for common vs rare variants (default=NULL). It should be a numeric value between 0 and 0.5, or NULL. When it is NULL, `1/ \sqrt{2 SampleSize }` will be used.
`test.type`	a string to indicate test type (default="Joint"). "Joint" indicates the joint test of the combined effects of common and rare variants. "Rare.Only" and "Common.Only" will conduct test only with rare and common variants, respectively.
`is_dosage`	see SKAT
`missing_cutoff`	see SKAT
`estimate_MAF`	see SKAT
`SetID1`	internal use only
`SSD.INFO`	an SSD_INFO object returned from Open_SSD.
`SetID`	a character value of Set ID. You can find a set ID of each set from SetInfo object of SSD.INFO
`SetIndex`	a numeric value of Set index. You can find a set index of each set from SetInfo object of SSD.INFO
`obj.SNPWeight`	output object from Read_SNP_WeightFile (default=NULL). If NULL, the beta weight with the “weights.beta” parameter will be used.
`...`	furthuer arguments to be passed to “SKAT_CommonRare”

Details

The small sample adjustment for binary traits is not implemented for "A" and "AR".

Value

`p.value`	p-value.
`p.value.resampling`	p-values from resampled phenotypes. You can get it when you use obj from SKAT_Null_Model function with resampling. See the SKAT_Null_Model.
`n.rare`	the number of rare variants used for the test
`n.common`	the number of common variants used for the test
`Cutoff`	the MAF cut-off to divide common and rare variants
`Q`	the test statistic of SKAT. It has NA when method="A" or "AR".
`param`	estimated parameters of each method.
`param$Is_Converged`	an indicator of the convergence. 1 indicates the method is converged, and 0 indicates the method is not converged. When 0 (not converged), "liu.mod" method is used to compute p-value.
`param$n.marker`	a number of SNPs in the genotype matrix
`param$n.marker.test`	a number of SNPs used for the test. It can be different from param$n.marker when some markers are monomorphic or have higher missing rates than the missing_cutoff.
`test.snp.mac`	a vector of minor allele count (MAC) of the snps tested. The name is SNP-ID.

Author(s)

Seunggeun Lee

References

Ionita-Laza, I.*, Lee, S.*, Makarov, V., Buxbaum, J. Lin, X. (2013). Sequence kernel association tests for the combined effect of rare and common variants. American Journal of Human Genetics, 92, 841-853. * contributed equally.

Examples



data(SKAT.example)
Z<-SKAT.example$Z


# continuous trait
obj<-SKAT_Null_Model(y.c ~ X, out_type="C", data=SKAT.example)
SKAT_CommonRare(Z, obj)$p.value
SKAT_CommonRare(Z, obj, method="A")$p.value
SKAT_CommonRare(Z, obj, method="AR")$p.value


# dichotomous trait 
obj<-SKAT_Null_Model(y.b ~ X, out_type="D", data=SKAT.example)

# Combined sum test in the manuscript (SKAT-C and Burden-C)
SKAT_CommonRare(Z, obj)$p.value
SKAT_CommonRare(Z, obj, r.corr.rare=1, r.corr.common=1 )$p.value


# Test only with common variant
SKAT_CommonRare(Z, obj, test.type="Common.Only")$p.value

# Test only with rare variant
SKAT_CommonRare(Z, obj, test.type="Rare.Only")$p.value


# Use CommonRare_Cutoff=0.01 instead of CommonRare_Cutoff = NULL
SKAT_CommonRare(Z, obj, CommonRare_Cutoff=0.01)$p.value

# Use custom weights; the first 10 variants have higher weights 
weights<-rep(1,67); weights[1:10]<-2
SKAT_CommonRare(Z, obj, weights=weights)$p.value

[Package SKAT version 2.2.5 Index]