Class ExomeDepth

Share:

Description

A class to hold the read count data that is used by ExomeDepth to call CNVs.

Objects from the Class

Objects can be created by calls of the form new("ExomeDepth", data = NULL, test, reference, formula = 'cbind(test, reference) ~ 1', subset.for.speed = NULL). data is optional and is only used if the formula argument refers to covariates (in which case these covariates must be included in the data frame). test and reference refer to the read count data for the test and reference samples. Creating a ExomeDepth object will automatically fit the beta-binomial model (using routines from the aod package) and compute the likelihood for the three copy number states (normal, deletion and duplication).

Slots

test:

numeric, read count data for the test sample.

reference:

numeric, read count data for the reference sample (usually a combination of samples).

formula:

character, a character string describing the linear model linking test and reference. Typically this would be cbind(test, reference) ~ 1.

expected:

The expected read count data for the test sample assuming normal copy number.

phi:

The over-dispersion parameter of the binomial model. See the aod package for more details.

likelihood:

A matrix of likelihood values, one column per copy number (deletion, normal, duplication).

annotations:

A data.frame specifying the chromosome, start and end for the bins used in the read count computation.

CNV.calls:

A data.frame describing the output of the CNV calling procedure.

Methods

CallCNVs

signature(x = "ExomeDepth", transition.probability = "numeric", chromosome = "factor", start = "numeric", end = "numeric", name = "character") ): Uses the pre-computed likelihood values and fits a hidden Markov Chain to the data to generated merged CNV calls.

AddAnnotations

signature(object = "ExomeDepth", name = "character", chromosome = "factor", start = "numeric", end = "numeric"): This method is unlikely to be directly used but it can include the exon names, chromosome, start, end into the ExomeDepth object.

TestCNV

signature(x = "ExomeDepth", chromosome = "factor", start = "numeric", end = "numeric", type = "character"): type must be either deletion of duplication. This function takes an ExomeDepth object and returns the Bayes factor in favor of a CNV at the specified location.

Author(s)

Vincent Plagnol

References

Paper recently submitted

See Also

select.reference.set CallCNVs aod

Examples

1
showClass("ExomeDepth")

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.