Segment.ASCAT: L2R and BAF joint segmentation using ASCAT.

In gustaveroussy/EaCoN: EaCoN : Easy Copy Number !

Description

This function performs the joint L2R and BAF segmentation using ASCAT PCF-based segmenter, centering and calling of WES, CytoScan, OncoScan or SNP6 data.

Usage

  Segment.ASCAT(data = NULL, mingap = 5E+06, smooth.k = NULL, BAF.filter = .75,
  homoCut = .05, penalty = 50, recenter = "l2r.centeredpeak",
  calling.method = "mad", nrf = 0.5, SER.pen = 40, out.dir = getwd(),
  return.data = FALSE, write.data = TRUE, plot = TRUE, force = FALSE)

Arguments

data	Data object contained in a RDS file generated with EaCoN.WES.Normalize() (for WES data), EaCoN.CS.Process() (for Affymetrix CytoScan data), EaCoN.OS.Process() (for Affymetrix OncoScan data), or EaCoN.SNP6.Process() (for Affymetrix GenomeWideSNP_6 data).
mingap	Minimum size (in nt) of a genomic gap (mainly, centromere), to allow ASCAT to treat flanking parts independently.
smooth.k	The half-window size for data smoothing by winsorization. If NULL, no winsorization performed.
BAF.filter	Perform a filtering of BAF noise, keeping a fraction of heterozygous BAF markers, represented by this parameter.
homoCut	Threshold value to consider a BAF segment as homozygote.
penalty	The penalty for ASPCF segmentation from ASCAT (see ?ASCAT::ascat.aspcf()).
recenter	Method to perform re-centering of the L2R profile. Can be : "l2r.mainpeak" (use the highest peak of the L2R density), "l2r.centeredpeak" (use the most-centered peak), "l2r.median" (just center to the median of the L2R distribution. Not recommended), any signed numeric (to force centering to a desired level), or NULL (do not recenter).
calling.method	Method to perform the calling of aberrations (attribution of the Loss / Normal / Gain status). Can be "mad" (build threshold based on the L2R spread noise, altered by the nrf coefficient) or "density" (automatically build thresholds based on the L2R density).
nrf	Coefficient to multiply the L2R spread noise, to build a calling threshold when calling.method is set to "mad". Not used when calling.method set to "density"
SER.pen	Penalty for the small events rescue step (a second-step segmentation of the L2R profile using PELT, to retrieve small events ignored by ASCAT). Can be a positive integer, or NULL (do not perform a second-step PELT segmentation at all).
out.dir	Output directory.
return.data	Return the results.
write.data	Write the results as a RDS file.
plot	Perform plots.
force	If some result files already exist, delete them before runing.

Details

Depending on the data source, different parameter values are recommended :

• for WES : SER.pen = 2

• for CytoScan array : SER.pen = 20, smooth.k = 5

• for OncoScan array : SER.pen = 40

• for SNP6 array : SER.pen = 60, smooth.k = 5, nrf = .25

SER.pen : The ASPCF algorithm from ASCAT::ascat.aspcf commonly ignores (very) small events (especially in loss) when not supported by the BAF profile). This option allows to detect such events using a PELT segmentation. See ?changepoint::cpt.mean().

Value

Several RDS objects, PNG plots, CBS files and data tables.

Author(s)

Bastien Job

gustaveroussy/EaCoN documentation built on Oct. 20, 2021, 2:41 a.m.