lva: lva

In pappewaio/AllelicImbalance: Investigates Allele Specific Expression

Description

make an almlof regression for arrays

Usage

lva(x, ...)

## S4 method for signature 'ASEset'
lva(
  x,
  rv,
  region,
  settings = list(),
  return.class = "LinkVariantAlmlof",
  type = "lm",
  verbose = FALSE,
  covariates = matrix(),
  ...
)

Arguments

x	ASEset object with phase and 'ref'/'alt' allele information
...	arguments to forward to internal functions
rv	RiskVariant object with phase and 'ref'/'alt' allele information
region	RiskVariant object with phase and alternative allele information
settings	RiskVariant object with phase and alternative allele information
return.class	'LinkVariantAlmlof' (more options in future)
type	"lm" or "nlme", "nlme" needs subject information
verbose	logical, if set TRUE, then function will be more talkative
covariates	add data.frame with covariates (only integers and numeric)

Details

internal method that takes one array with results from regionSummary and one matrix with group information for each risk SNP (based on phase)

Author(s)

Jesper R. Gadin, Lasse Folkersen

Examples

data(ASEset) 
a <- ASEset
# Add phase
set.seed(1)
p1 <- matrix(sample(c(1,0),replace=TRUE, size=nrow(a)*ncol(a)),nrow=nrow(a), ncol(a))
p2 <- matrix(sample(c(1,0),replace=TRUE, size=nrow(a)*ncol(a)),nrow=nrow(a), ncol(a))
p <- matrix(paste(p1,sample(c("|","|","/"), size=nrow(a)*ncol(a), replace=TRUE), p2, sep=""),
	nrow=nrow(a), ncol(a))

phase(a) <- p

#add alternative allele information
mcols(a)[["alt"]] <- inferAltAllele(a)

#init risk variants
p.ar <- phaseMatrix2Array(p)
rv <- RiskVariantFromGRangesAndPhaseArray(x=GRvariants, phase=p.ar)

#colnames has to be samea and same order in ASEset and RiskVariant
colnames(a) <- colnames(rv)

# in this example each and every snp in the ASEset defines a region
r1 <- granges(a)

#use GRangesList to merge and use regions defined by each element of the
#GRangesList
r1b <- GRangesList(r1)
r1c <- GRangesList(r1, r1)

# in this example two overlapping subsets of snps in the ASEset defines the region
r2 <- split(granges(a)[c(1,2,2,3)],c(1,1,2,2))

# link variant almlof (lva)
lva(a, rv, r1)
lva(a, rv, r1b)
lva(a, rv, r1c)
lva(a, rv, r2)

# Use covariates (integers or nuemric)
cov <- data.frame(age=sample(20:70, ncol(a)), sex=rep(c(1,2), each=ncol(a)/2),  
row.names=colnames(a))
lva(a, rv, r1, covariates=cov)
lva(a, rv, r1b, covariates=cov)
lva(a, rv, r1c, covariates=cov)
lva(a, rv, r2, covariates=cov)

# link variant almlof (lva), using nlme
a2 <- a
ac <- assays(a2)[["countsPlus"]]
jit <- sample(c(seq(-0.10,0,length=5), seq(0,0.10,length=5)), size=length(ac) , replace=TRUE)
assays(a2, withDimnames=FALSE)[["countsPlus"]] <- round(ac * (1+jit),0)
ab <- cbind(a, a2)
colData(ab)[["subject.group"]] <- c(1:ncol(a),1:ncol(a))
rv2 <- rv[,c(1:ncol(a),1:ncol(a))]
colnames(ab) <- colnames(rv2)

lva(ab, rv2, r1, type="nlme")
lva(ab, rv2, r1b, type="nlme")
lva(ab, rv2, r1c, type="nlme")
lva(ab, rv2, r2, type="nlme")

pappewaio/AllelicImbalance documentation built on April 11, 2020, 2:58 a.m.