calmateByTotalAndFracB.array {calmate}

Normalize allele-specific copy numbers (total,fracB)

Description

Normalize allele-specific copy numbers (total,fracB), where total is the total (non-polymorphic) signal and fracB is the allele B fraction. It is only loci with a non-missing (NA) fracB value that are considered to be SNPs and normalized by CalMaTe. The other loci are left untouched.

Usage

## S3 method for class 'array'
calmateByTotalAndFracB(data, references=NULL, ..., refAvgFcn=NULL, verbose=FALSE)

Arguments

data

An Jx2xI numeric array, where J is the number of loci, 2 is total and fracB (in that order, if unnamed), and I is the number of samples.

references

A logical or numeric vector specifying which samples should be used as the reference set. By default, all samples are considered. If not NULL at least 3 samples.

...

Additional arguments passed to *calmateByThetaAB().

refAvgFcn

(optional) A function that takes a JxI numeric matrix an argument na.rm and returns a numeric vector of length J. It should calculate some type of average for each of the J rows, e.g. rowMedians. If specified, then the total copy numbers of the calibrated ASCNs are standardized toward (twice) the average of the total copy numbers of the calibrated reference ASCNs.

verbose

See Verbose.

Value

Returns an Jx2xI numeric array with the same dimension names as argument data.

References

[1] M. Ortiz-Estevez, A. Aramburu, H. Bengtsson, P. Neuvial and A. Rubio, CalMaTe: A method and software to improve allele-specific copy number of SNP arrays for downstream segmentation, Bioinformatics, 2012 [PMC3381965].

See Also

To calibrate (thetaA,thetaB) or (CA,CB) signals, see *calmateByThetaAB().

Examples

# Load example (thetaA,thetaB) signals
path <- system.file("exData", package="calmate");
theta <- loadObject("thetaAB,100x2x40.Rbin", path=path);

# Transform to (total,fracB) signals
data <- thetaAB2TotalAndFracB(theta);

# Calibrate (total,fracB) by CalMaTe
dataC <- calmateByTotalAndFracB(data);

# Calculate copy-number ratios
theta <- data[,"total",];
thetaR <- matrixStats::rowMedians(theta, na.rm=TRUE);
data[,"total",] <- 2*theta/thetaR;

# Plot two "random" arrays
Clim <- c(0,4);
Blim <- c(0,1);
subplots(4, ncol=2, byrow=FALSE);
for (ii in c(1,5)) {
  sampleName <- dimnames(data)[[3]][ii];
  sampleLabel <- sprintf("Sample #%d ('%s')", ii, sampleName);
  plot(data[,,ii], xlim=Clim, ylim=Blim);
  title(main=sampleLabel);
  plot(dataC[,,ii], xlim=Clim, ylim=Blim);
  title(main=sprintf("%s\ncalibrated", sampleLabel));
}


# Assert that it also works with a single unit
dummy <- calmateByTotalAndFracB(data[1,,,drop=FALSE]);
stopifnot(length(dim(dummy)) == 3);