traseR: TRait-Associated SNP EnRichment analyses

In traseR: GWAS trait-associated SNP enrichment analyses in genomic intervals

Description

Perform GWAS trait-associated SNP enrichment analyses in genomic intervals using different approaches

Usage

traseR(snpdb, region, snpdb.bg=NULL, keyword = NULL, rankby = c("pvalue", "odds.ratio"), 
test.method = c("binomial", "fisher","chisq", "nonparametric"), alternative = c("greater", "less", "two.sided"), 
ntimes=100,nbatch=1,
trait.threshold = 0, traitclass.threshold=0, pvalue = 1e-3)

Arguments

snpdb	A GRange object. It could be GWAS trait-associated SNPs downloaded from up-to-date dbGaP and NHGRI public database. It is maintained to be updated to the latest version. The data frame contains the following columns,Source,Trait,SNP,p.value,Chr,Position,Context,GENE_NAME,GENE_START,GENE_END,GENE_STRAND. The data frame is in data subdirectory. Users are free to add more SNP records to the data frame for practical use. It could also be a data frame with columns as, SNP,Chr,Position.
region	A GRange object or data frame, which is genomic intervals with three columns, chromosome, genomic start position, genomic end position.
snpdb.bg	A GRange object contains non-trait-associated SNPs. They are treated as background for statistical testing instead of whole genome as background if specified.
keyword	The keyword is used when specific trait is of interest. If keyword is specified, only the SNPs associated to the trait are used for analyses. Otherwise, all traits will be analyzed.
rankby	Traits could be ranked by either p-value or adds.ratio based on the enrichment level of trait-associated SNPs in genomic intervals.
test.method	Several hypothesis testing options are provided: binomial(binomial test),fisher(Fisher's exact test),chisq(Chi-squared test),chisq(nonparametric test). Default is binomial(binomial test)
alternative	Indicate the alternative hypothesis. If greater, test if the genomic intervals are enriched in trait-associated SNPs than background. If less, test if the genomic intervals are depleted in trait-associated SNPs than background. If two.sided, test if there is difference between the enrichment of trait-associated SNPs in genomic intervals and in background.
ntimes	The number of shuffling time for one batch. See nbatch.
nbatch	The number of batches. The product of ntimes and nbatch is the total number of shuffling time.
trait.threshold	Test traits with number of SNPs more than the threshold.
traitclass.threshold	Test trait class with number of SNPs more than the threshold.
pvalue	SNPs with p-value less than this threshold are used for analyses.

Details

Return a list that contains three data frames. One data frame tb.all contains the results of enrichment analyses for all trait-associated SNPs in genomic intervals. Another data frame tb1 contains the results of enrichment analyses for each trait-associated SNPs in genomic intervals separately. Another data frame tb2 contains the results of enrichment analyses for each trait-class-associated SNPs in genomic intervals separately.

Value

The data frame tb1 has columns,

Trait	Name of trait
p.value	P-value calculated from hypothesis testing
q.value	Adjusted p-value from multiple testing using FDR correction
odds.ratio	Odds ratio calculated based on number of trait-associated SNPs in genomic intervals, number of trait-associated SNPs across whole genome, genomic intervals size (bps) and genome size (bps)
taSNP.hits	Number of trait-associated SNPs in genomic intervals
taSNP.num	Number of SNPs for specific trait

Author(s)

Li Chen <[e-mail]>, Zhaohui S.Qin<[e-mail]>

See Also

print.traseR

Examples

	data(taSNP)
	data(Tcell)
	x=traseR(snpdb=taSNP,region=Tcell)
	print(x)

traseR documentation built on Nov. 8, 2020, 6:55 p.m.