burdenMeta: Meta analyze burden tests from multiple studies
In seqMeta: Meta-Analysis of Region-Based Tests of Rare DNA Variants

Description Usage Arguments Details Value Author(s) See Also Examples

Takes as input 'seqMeta' objects (from the prepScores function), and meta-analyzes the corresponding burden test.

1 2	burdenMeta(..., SNPInfo = NULL, wts = 1, snpNames = "Name", aggregateBy = "gene", mafRange = c(0, 0.5), verbose = FALSE)

`...`	`seqMeta` objects
`SNPInfo`	The SNP Info file. This should contain the fields listed in snpNames and aggregateBy. Only SNPs in this table will be meta analyzed, so this may be used to restrict the analysis.
`wts`	weights for the burden test, as a function of maf, or a character string specifying weights in the SNP Info file.
`snpNames`	The field of SNPInfo where the SNP identifiers are found. Default is 'Name'
`aggregateBy`	The field of SNPInfo on which the skat results were aggregated. Default is 'gene'. Though gene groupings are not explicitely required for single snp analysis, it is required to find where single snp information is stored in the seqMeta objects.
`mafRange`	Range of MAF's to include in the analysis (endpoints included). Default is all SNPs (0 <= MAF <= 0.5).
`verbose`	logical. Whether progress bars should be printed.

This function uses the scores and their variances available in a seqMeta object to perform burden tests. Though coefficients are reported, the tests are formally score tests, and the coefficients can be thought of as one-step approximations to those reported in a Wald test.

a data frame with the following columns:

`gene`	the name of the gene or unit of aggregation being meta analyzed
`p`	the p-value from the burden tests
`beta`	approximate coefficient for the effect of genotype
`se`	approximate standard error for the effect of genotype
`cmafTotal`	the cumulative minor allele frequency of the gene
`cmafUsed`	the cumulative minor allele frequency of snps used in the analysis
`nsnpsTotal`	the number of snps in the gene
`nsnpsUsed`	the number of snps used in the analysis
`nmiss`	The number of 'missing' SNPs. For a gene with a single SNP this is the number of individuals which do not contribute to the analysis, due to studies that did not report results for that SNP. For a gene with multiple SNPs, is totalled over the gene.

Arie Voorman, Jennifer Brody

skatMeta skatOMeta prepScores

###load example data for two studies:
### see ?seqMetaExample	
data(seqMetaExample)

####run on each cohort:
cohort1 <- prepScores(Z=Z1, y~1, SNPInfo=SNPInfo, data=pheno1)
cohort2 <- prepScores(Z=Z2, y~1, SNPInfo=SNPInfo, data=pheno2)

#### combine results:
out <- burdenMeta(cohort1, cohort2, SNPInfo = SNPInfo, mafRange=c(0,.01))
head(out)

## Not run: 
##### Compare with analysis on full data set:
bigZ <- matrix(NA,2*n,nrow(SNPInfo))
colnames(bigZ) <- SNPInfo$Name

for(gene in unique(SNPInfo$gene)) {
   snp.names <- SNPInfo$Name[SNPInfo$gene == gene]
     bigZ[1:n,SNPInfo$gene == gene][ , snp.names \%in\% colnames(Z1)] <- 
                   Z1[ , na.omit(match(snp.names,colnames(Z1)))]
     bigZ[(n+1):(2*n),SNPInfo$gene == gene][ , snp.names \%in\% colnames(Z2)] <- 
                   Z2[ , na.omit(match(snp.names,colnames(Z2)))]
}

pheno <- rbind(pheno1[, c("y","sex","bmi")],pheno2[,c("y","sex","bmi")])
burden.p <- c(by(SNPInfo$Name, SNPInfo$gene, function(snp.names) {
   inds <- match(snp.names,colnames(bigZ)) burden <- rowSums(bigZ[,na.omit(inds)],na.rm=TRUE)
   mod <- lm(y~burden + gl(2,nrow(pheno1)),data=pheno)
   summary(mod)$coef[2,4]
}))

head(cbind(out$p,burden.p))

#will be slightly different:
plot(y=out$p,x=burden.p, ylab = "burden meta p-values", xlab = "complete data p-values")

## End(Not run)