| rnb.execute.pOOBAH {RnBeads} | R Documentation |
Probe signal intensities are masked based on their out-of-band signal intensities to counter hybridization failure.
rnb.execute.pOOBAH( raw.set, anno.table = NULL, pval.thresh = 0.05, verbose = FALSE )
raw.set |
Methylation dataset as an instance of |
anno.table |
Annotation for |
pval.thresh |
Computed detection p-values above this threshold are masked. Default value is 0.05. |
verbose |
If set to true, a short information is printed on how many probes are masked by the method. |
rnb.execute.pOOBAH is used to apply the method pOOBAH (P-value with OOB probes for Array Hybridization),
which was conceived by Zhou, Triche, Laird and Shen to mask probes associated with hybridization failures.
pOOBAH has been implemented in the R-package "sesame", a dependency needed for this function (see Zhou et al, 2018 and the respective Bioconductor/github pages).
pOOBAH computes the detection p-values by constructing 2 empirical cumulative density functions (eCDFs) based
on the out-of-band signal intensities of the red and the green channel, respectively, to detect hybridization failures.
The (in-band) green and red channel signal intensities of the probes are passed to the eCDFs and the probes with a
p-value higher than the given threshold (pval.thresh) are masked, as they are considered background.
pOOBAH is applied separately to each sample.
Hybridization failures might occur due to somatic or germline deletions.
In addition, unreliable low-intensity probes might also be masked.
Returns a modified RnBeadRawSet, in which signal intensities are masked, if their computed p-value
was greater than pval.thresh. Note, in datasets with several samples, signal intensities of a specific probe
might be masked in sample A, but not in sample B, as pOOBAH is applied separately to each sample.
For example: the signal intensities of probe cg24488772 might be masked in sample 1, but not in sample 12.
pOOBAH method: Wanding Zhou. Adapted by Nathan Steenbuck.
library(RnBeads.hg19) data(small.example.object) filtered <- rnb.execute.pOOBAH(rnb.set.example)