model.apply: Computes copy number for a set of CGH segments

In cghRA: Array CGH Data Analysis and Visualization

Description

This function translates log ratios of a set of segments into copy numbers, applying a copy number model as produced by model.auto or model.test.

If exact is set set to FALSE, copy numbers are rounded and consecutive segments with the same copy number are merged.

Usage

  model.apply(segStarts, segEnds, segChroms, segLogRatios, segLengths, model = NA,
    center = model['center'], width = model['width'], ploidy = model['ploidy'],
    exact = FALSE, merge = TRUE)

Arguments

segStarts	Numeric vector, the starting positions of the CGH segments to modelize.
segEnds	Numeric vector, the endind positions of the CGH segments to modelize.
segChroms	Vector, the chromosome holding the CGH segments to modelize.
segLogRatios	Double vector, the log ratios of the CGH segments to modelize.
segLengths	Numeric vector, the lengths of the CGH segments to modelize.
model	A numeric vector, as returned by model.auto or model.test. Can be NA if parameters are provided via other arguments.
center	Single double value, the center parameter to use in the model.
width	Single double value, the width parameter to use in the model.
ploidy	Single numeric value, copy number supposed to be the most common within the analyzed genome.
exact	Single logical value, whether to return continue copy numbers (double) or discrete ones (integer).
merge	Single logical value, whether to merge consecutive segments with the same copy number when exact is FALSE.

Value

Returns a data.frame describing the segments :

segStarts	Extracted from the segStarts argument.
segEnds	Extracted from the segEnds argument.
segChroms	Extracted from the segChroms argument.
segLogRatios	Double, the theoretic log ratio of the segment, with 2 copies as reference.
segCopies	Numeric, the copy number of the segment.
segLengths	Extracted from the segLengths argument.

Author(s)

Sylvain Mareschal

See Also

copies, model.auto, model.test

Examples

  # Generating random segmentation results
  ## with 30% normal cells contamination
  ## with +10% for normal DNA labelling
  segLogRatios <- c(
    rnorm(
      sample(5:20, 1),
      mean = log((1*0.7 + 2*0.3)/(2*1.1), 2),   # One deletion
      sd = 0.08
    ),
    rnorm(
      sample(80:120, 1),
      mean = log(2/(2*1.1), 2),                 # No alteration
      sd = 0.08
    ),
    rnorm(
      sample(40:60, 1),
      mean = log((3*0.7 + 2*0.3)/(2*1.1), 2),   # One more copy
      sd = 0.08
    )
  )
  segLogRatios <- sample(segLogRatios)
  segLengths <- as.integer(3 + round(rchisq(length(segLogRatios), 1)*100))
  segEnds <- cumsum(segLengths)
  segStarts <- c(1L, head(segEnds, -1))
  segChroms <- rep("chr1", length(segEnds))
  
  # Generated genome
  genome <- data.frame(
    segChroms,
    segStarts,
    segEnds,
    segLogRatios,
    segLengths
  )
  print(genome)
  
  # Automatic modelization
  model <- model.auto(
    segLogRatios = segLogRatios,
    segChroms = segChroms,
    segLengths = segLengths
  )
  
  # Profile simplification
  segments <- model.apply(
    segStarts,
    segEnds,
    segChroms,
    segLogRatios,
    segLengths,
    model = model,
    exact = FALSE,
    merge = TRUE
  )
  layout(matrix(1:2, ncol=1))
  plot(x=segStarts, y=segLogRatios, type="s", xlab="Position", ylab="Log Ratios")
  plot(x=segments$segStarts, y=segments$segCopies, type="s", xlab="Position", ylab="Copies")
  print(segments)
  
  layout(1)

cghRA documentation built on May 2, 2019, 3:34 a.m.