selectRegions {bayesCGH}R Documentation

Selects regions with high proportions of chromosomal gains and losses.

Description

The 'selectRegions' function takes a 'ClassifySegList' object (or subset) and identifies those regions on the genome that show high likelihoods of high proportions of chromosomal gains or losses.

The 'selectAmpsDels' function simply selects those regions on the genome that have previously been classified as amplified or deleted.

Usage

selectRegions(seg, arrays, plot = FALSE, summarise = TRUE,
                  prob.prop = 0.9, prop.greater.than = 0.2, group.name =
"", ...)

selectAmpsDels(seg, arrays, group.name = "")

Arguments

seg Object of class ClassifySegList
arrays Vector describing which arrays of 'seg' should be considered in region selection.
plot Should the selected regions be plotted (plotProportions)?
summarise Should the selected regions be summarised (see Details).
prob.prop Threshold probability; see Details.
prop.greater.than Threshold proportion; see Details.
group.name Name for group of arrays selected in 'arrays'.
... Additional arguments to be passed to plotProportions.

Details

This function takes a 'ClassifySegList' object (or subset) and identifies those regions on the genome that show high likelihoods of high proportions of chromosomal gains or losses. Specifically, given a threshold proportion 'prop.greater.than', empirical Bayes methods are applied to calculate the posterior probability that the proportion of chromosomal gains or losses exceeds this threshold for each probe on the arrays.

Probe selection is then made on the basis of those regions for which this posterior probability exceeds some probability threshold 'prob.prop'.

If 'summarise = TRUE' the selected probes are summarised into continuous regions (summariseRegions).

Value

List containing:

data Description of data used.
gains Probability for each probe that the proportion of samples with a chromosomal gain at that probe exceeds that given as 'prop.greater.than'.
losses Probability for each probe that the proportion of samples with a chromosomal loss at that probe exceeds that given as 'prop.greater.than'.
summary If 'summarise = TRUE', a summary of regions (and genes) that have a probability greater than 'prob.prop' of having a proportion of samples with a chromosomal gain or loss exceeding that of 'prop.greater.than'.

Author(s)

Thomas Hardcastle

References

Hardcastle 2008

See Also

nudSegmentation


[Package bayesCGH version 0.6 Index]