Skip to contents

Try to find a better peptide-spectrum match than the current PSM.

Usage

findBetterMatch(
  x,
  peptideVariable = "peptide",
  modificationsList = NULL,
  maxMods = 2,
  improvedBy = c("maxScore", "maxCount"),
  convertToStyle = "name"
)

Arguments

x

Spectra() object containing identifications.

peptideVariable

character(1L) pointing to the peptide sequence variable in x.

modificationsList

character() with all modifications to try out. The modifications need to be present in the PTMods::modifications dataset from the unimod database. Note that the more modifications are included, the more combinations are possible and the longer it takes to run.

maxMods

integer(1L) indicating the maximum modifications allowed at the same time on one sequence.

improvedBy

character(1L): either "maxScore" or "maxCount", defining how to choose the better match. "maxScore" keeps the match with the highest ndotproduct and "maxCount" keeps the one with the most matched peaks.

convertToStyle

character(1L) to choose from those available in PTMods::convertAnnotation(). By default, the unimod "name" is given.

Value

A named character() vector of length equal to the number of spectra with identifications. The values are the best matches found using the modificationsList and the names are the indices for those sequences in the innitial Spectra object x.

Details

This function is meant to help finding a better peptide-spectrum match for identifications that fail certain checks (i.e. checkOverlap() has returned false) or if the PSM's FDR doesn't qualify it as a confident identification. A peptide sequence is returned (with or without modifications) for which the greatest number of matched peaks/a better ndotproduct (see MsCoreUtils::ndotproduct() for more details) was obtained.

Author

Guillaume Deflandre

Examples

if (FALSE) { # \dontrun{
## Load the data: a Spectra object with identifications
data(spBoekweg)
data(psmBoekweg)
head(psmBoekweg$pkey <- paste0(basename(psmBoekweg$filename),
                               sub("^.+scan=", "::", psmBoekweg$scannr)))
head(spBoekweg$pkey <- paste0(basename(spBoekweg$dataOrigin),
                              sub("^.+scan=", "::", spBoekweg$spectrumId)))
sp <- Spectra::joinSpectraData(spBoekweg, psmBoekweg, by.x = "pkey")

## Run with stock modifications (see PSMatch:::.ptmNames)
findBetterMatch(sp[30], "peptide", maxMods = 1)

## Run with custom modifications (need to be present in
## `modifications$Name`). Notice that for this example, the identified match
## includes an oxidation. We are removing it to show that findBetterMatch()
## also identifies the same sequence.
sp[30]$sequence <- getCanonicalSequence(sp[30]$peptide)
findBetterMatch(sp[30], "sequence", maxMods = 1)

## Get the best match based on the number of matched peaks
findBetterMatch(sp[30], "sequence", maxMods = 1, improvedBy = "maxCount")
} # }