runSegmentation: Run up to 4 different copy number segmentation algorithms.

Description Usage Arguments Details Value Author(s) Examples

View source: R/runSegmentation.R

Description

runSegmentation Segments binned and normalized copy number values using multiple algorithms and allows the summarization of the segmentation results. Automatically calculates the "gain", "loss" calls for each segment.

Usage

1
2
3
4
5
6
7
runSegmentation(x, isLogTransformed = TRUE, genome = NULL,
  segmentType = c("CBS", "LACBS", "HMM", "PLS"), summaryMethod = "mean",
  segmentsToSummarise = segmentType, callTypeLog2R = "summary",
  callThreshLog2R = c(-0.15, 0.15), minimumBinsPerSegment = 3,
  gapLength = 1e+06, normalizeSegmentedBins = TRUE, inter = c(-0.1, 0.1),
  alpha = 1e-10, undo.splits = "sdundo", undo.SD = 1,
  segmentStatistic = "seg.mean")

Arguments

x

Either a CNAclinicData object from the output of processForSegmentation function or a data.frame with specific columns (see details) or a QDNAseqCopyNumbers object.

isLogTransformed

if TRUE (default) data is assumed to be log2 transformed. If FALSE, transformation is carried out prior to segmentation.

genome

Genome build used to align sequencing reads.

segmentType

One or more algorithms for segmentation. i.e. c("CBS", "HMM", "PLS", "LACBS")

summaryMethod

Summarization method for ensemble segmentation. i.e. One of c("mean", "median", "min", "max", "Q1", "Q3")

segmentsToSummarise

X

callTypeLog2R

Segment type used to call CNAs (default "summary")

callThreshLog2R

Thresholds used to call segments as a "loss" or "gain". Defaults to c(-0.15, 0.15)

minimumBinsPerSegment

Minimum number of bins in each segment, default is 3. Argument is specific to PLS algorithm.

gapLength

Minimum length in base-pairs between the two closest loci to consider a region to be a "gap". (default 1000000) Argument is specific to LACBS algorithm.

normalizeSegmentedBins

Normalizes the segmented bins if TRUE (default).

inter

The interval in which the function should search for the normal level. Utilised in normalizeSegmentedBins.

alpha

Significance levels for the test to accept change-points. Default is 1e-10

undo.splits

A character string specifying how change-points are to be undone, if at all. Default is "sdundo", which undoes splits that are not at least this many SDs apart. Other choices are "prune", which uses a sum of squares criterion, and "none".

undo.SD

The number of SDs between means to keep a split if undo.splits="sdundo". Default is 1.0.

segmentStatistic

Default is "seg.mean".

Details

If x is a data.frame it should contain the columns: chromosome, start, end and the optional column usebin followed by separate columns named after each different sample contaning normalized and log-transformed copy number measurements. chromosome should be a character vector containing only 1-22, X, Y while all other columns should contain numerical values. The genomic coordinates should belong to fixed-width bins.

Value

Returns an object of class CNAclinicData with segmentation results from the chosen algorithms as well as the summarised result.

Author(s)

Dineika Chandrananda

Examples

1
2
3
4
     ## Not run: 
      vignette("CNAclinic")
     
## End(Not run)

sdchandra/CNAclinic documentation built on May 29, 2019, 9:33 a.m.