align.fsa: Aligns peaks using size ladder

In MLPA: Multiplex Ligation-Dependent Probe Amplification Data Analysis

Description

This function adds to a fsa object a linear regression model allowing the raw time indexes to be converted into base pair sizes, using a known size markers ladder.

Usage

  align.fsa(x, channel = "ROX", fullLadder = c(50, 60, 90, 100, 120, 150, 160, 180, 190,
    200, 220, 240, 260, 280, 290, 300, 320, 340, 360, 380, 400), useLadder = c(50, 60, 90,
    100, 120), outThreshold = 0.15, noiseLevel = 10, surePeaks = 5,
    trim = c("forward", "backward", "none"), maskOffScale = FALSE, rMin = 0.999,
    rescue = FALSE, ylim = NULL, ...)

Arguments

x	An object of class fsa, as returned by read.fsa
channel	Single character value, the name of the channel used for size markers.
fullLadder	Integer vector, the size markers used in the assay (in base pairs).
useLadder	Integer vector, the size markers to use for the alignment (using only size markers nearing the expected size for the experimental peaks may achieve a more precise alignment). They must be present in fullLadder. If NULL, fullLadder will be used entirely.
outThreshold	Single numeric value, maximal distance from the computed size-marker intensity for a peak to be considered as a size-marker. If lower than 1, it is considered as a proportion of the size-marker intensity computed from sure peaks.
noiseLevel	Single numeric value, minimal intensity for a local maximum to be considered as a peak.
maskOffScale	Single logical value, whether to mask indexes with off-scale values in any channel to limit side-effects or not.
surePeaks	Single integer value, amount of peaks to use to compute size-marker intensity. They are selected at the end of the profile, as most artefacts are observed ate the beginning. Consider to reduce this value if your assay was prematurely ended.
trim	Single character value, defining how to behave when more/less peaks than expected are read. "forward" will keep first peaks and adjust discarding the last ones, "backward" will keep last peaks and adjust discarding the first ones, and "none" will generate an error.
rMin	Single numeric value, minimum adjusted r squared value (see summary.lm) to consider an alignment as "good". Poor alignments raise a warning, and may be due to artefactual peaks in the size-marker channel or errors in fullLadder definition. Consider lowering outThreshold and raising noiseLevel to minimize artefact selection.
rescue	Single logical value, whether to plot a "rescue" profile or not. Rescue profiles are calls to plot.fsa on which diverse additionnal data is drawn to help diagnose alignment problems.
ylim	To be passed to plot.fsa for the alignment rescue plot, if enabled (see rescue).
...	Further arguments to be passed to plot.fsa for the alignment rescue plot, if enabled (see rescue).

Value

Returns the object of class fsa provided with updated attributes :

ladderModel	A numeric vector of linear regression coefficients to use to convert raw indexes into base pairs.
ladderExact	A named numeric vector of raw indexes at which size markers were detected.

Author(s)

Sylvain Mareschal, Philippe Ruminy

Examples

  # Example FSA file provided
  fsa <- read.fsa(system.file("extdata/fsa/A5918.fsa", package="MLPA"))
  
  # Plot subset of the profile (time index)
  plot(fsa, units="index", xlim=c(4000,5000))
  
  # Align using full ladder
  fullLadder <- c(
    50, 60, 90, 100, 120, 150, 160, 180, 190, 200, 220,
    240, 260, 280, 290, 300, 320, 340, 360, 380, 400
  )
  fsa <- align.fsa(fsa, fullLadder=fullLadder)
  
  # Plot subset of the profile (base pairs)
  plot(fsa, units="bp", xlim=c(80,130))

MLPA documentation built on May 2, 2020, 1:06 a.m.