burden.subscores: Linear, logistic or multinomial regression on a multiple...

In Ravages: Rare Variant Analysis and Genetic Simulations

Linear, logistic or multinomial regression on a multiple genetic scores within a genomic region

Description

Performs burden tests with subscores in the regression on categorical or continuous phenotypes

Usage

burden.subscores(x, NullObject, genomic.region = x@snps$genomic.region, 
                 SubRegion = x@snps$SubRegion, burden.function = WSS, 
                 maf.threshold = 0.5, get.effect.size = FALSE, 
                 alpha = 0.05, cores = 10, verbose = TRUE)

Arguments

x	A bed matrix
NullObject	A list returned from NullObject.parameters
genomic.region	A factor containing the genomic region of each SNP, x@snps$genomic.region by default, for example the CADD regions
SubRegion	A vector containing subregions within each genomic.region, x@snps$SubRegion by default, for example genomic categories
burden.function	A function to compute the genetic score, WSS by default.
maf.threshold	The MAF threshold to use for the definition of a rare variant in the CAST score. Set at 0.5 by default
get.effect.size	TRUE/FALSE: whether to return effect sizes of the tested genomic.region (OR for categorical phenotypes, betas for continuous phenotypes)
alpha	The alpha threshold to use for the OR confidence interval
cores	How many cores to use, set at 10 by default. Only needed if NullObject$pheno.type = "categorical"
verbose	Whether to display information about the function actions

Details

This function will return results from the regression of the phenotype on the genetic score(s) for each genomic region. Within each genomic region, a subscore will be computed for each SubRegion and one test will be performed for each genomic.region.

When used after set.CADDregions, it will perform a test by CADD region with one subscore by genomic category (coding, regulatory, intergenic) as in the RAVA.FIRST() strategy.

If only two groups of individuals are present, a classical logistic regression is performed. If more than two groups of individuals are present, a non-ordinal multinomial regression is performed, comparing each group of individuals to the reference group indicated by the argument ref.level in NullObject.parameters. The choice of the reference group won't affect the p-values, but only the Odds Ratios. In both types of regression, the p-value is estimated using the Likelihood Ratio test and the function burden.mlogit.

If the phenotype is continuous, a linear regression is performed using the function burden.continuous.

The type of phenotype is determined from NullObject$pheno.type.

Value

A dataframe with one row per genomic region and two columns:

p.value	The p.value of the regression
is.err	0/1: whether there was a convergence problem with the regression

If get.effect.size=TRUE, a list is returned with the previous dataframe in $Asso and with effect, a list containing matrices with three columns:

OR/beta	The OR/beta value(s) associated to the subscores in the regression. For categorical phenotypes, if there are more than two groups, there will be one OR value per group compared to the reference group
l.lower	The lower bound of the confidence interval of each OR/beta
l.upper	The upper bound of the confidence interval of each OR/beta

See Also

RAVA.FIRST, NullObject.parameters, burden.continuous.subscores, burden.mlogit.subscores, CAST, WSS

Examples

#Import 1000Genome data from region around LCT gene
x <- as.bed.matrix(x=LCT.matrix.bed, fam=LCT.matrix.fam, bim=LCT.snps)

#Group variants within known genes and
#Within coding and regulatory regions
x <- set.genomic.region.subregion(x, regions = genes.b37, 
                                  subregions = subregions.LCT)

#Add population
x@ped[,c("pop", "superpop")] <- LCT.matrix.pop1000G[,c("population", "super.population")]

#Select EUR superpopulation
x <- select.inds(x, superpop=="EUR")
x@ped$pop <- droplevels(x@ped$pop)

#Keep only variants with a MAF lower than 1%
x1 <- filter.rare.variants(x, filter = "whole", maf.threshold = 0.01)

#run null model, using the 1000Genome population as "outcome"
x1.H0 <- NullObject.parameters(pheno = x1@ped$pop, ref.level = "CEU",
                               RVAT = "burden", pheno.type = "categorical")

#run functionally-informed burden test WSS in LCT
burden.subscores(select.snps(x1, genomic.region == "LCT"), 
                 NullObject = x1.H0, burden.function = WSS, 
                 get.effect.size=FALSE, cores = 1)


####Using the RAVA-FIRST approach with CDD regions
#Group variants within CADD regions and genomic categories
#x <- set.CADDregions(x)

#Filter of rare variants: only non-monomorphic variants with
#a MAF lower than 2.5%
#and with a adjusted CADD score greater than the median
#x1 <- filter.adjustedCADD(x, filter = "whole", maf.threshold = 0.025)

#run functionally-informed burden test WSS
#burden.subscores(x1, NullObject = x1.H0, burden.function = WSS, 
#                 get.effect.size=FALSE, cores = 1)

Ravages documentation built on April 1, 2023, 12:08 a.m.