filterBySibPepCorr: Filter by sibling peptide correlation

In CCprofiler/CCprofiler: Complex feature detection using SEC protein chromatograms.

Description

Filter peptides in traces object based on Sibling Peptide Correlation (SibPepCorr). Since peptides belonging to the same protein should co-elute in a PCP experiment the quality of their signal can be estimated using the average correlation to its sibling peptides. Additionally comparing target and decoy peptides allows for robust FDR estimation.

Usage

filterBySibPepCorr(traces, fdr_cutoff = 0.01, fdr_type = "protein",
  FFT = 1, absolute_spcCutoff = NULL, rm_decoys = FALSE,
  plot = TRUE, PDF = FALSE, CSV = FALSE)

Arguments

traces	An object of class traces (the trace_type must be "peptide").
fdr_cutoff	Numeric, specifying the FDR cutoff to be applied to select a sibling peptide correlation cutoff
FFT	Numeric, specifying the fraction of false targets (FFT). Default is 1.
absolute_spcCutoff	Numeric, specifying an absolute sibling correlation cutoff to be applied. Only used if fdr_cutoff is not provided. Default is NULL.
rm_decoys	Logical, specifying if decoys should be removed. It is strongly recommended to keep the decoys, for visual assessment. Default is FALSE.
plot	Logical, wether to print Score distribution and precision-recall plots to R console.
PDF	Logical, wether to print Score distribution and precision-recall plots to "SibPepCorrFilter_IDFDRplot.pdf" in working directory
CSV	Logical, write a table output of the FDR estimation results to the working directory

Value

a traces object, only containing peptides passing the specified spc/fdr cutoff. If the input traces object does not contain calculated spc values for every peptide a SibPepCorr column is added.

Examples

# Load example data
tracesRaw <- examplePeptideTraces

## Filter the raw traces object to a protein FDR of 1%
  tracesFiltered <- filterBySibPepCorr(traces = tracesRaw,
                                       fdr_cutoff = NULL,
                                       fdr_type = "protein",
                                       FFT = 1,
                                       absolute_spcCutoff = 0.2,
                                       rm_decoys = FALSE,
                                       plot = TRUE,
                                       CSV = FALSE)
  ## Compare the filtered traces to the raw traces
summary(tracesRaw)
summary(tracesFiltered)
##--------------------------------------------------------------------------------
## Alternatively the Sibling peptide correlation can also be calculated seperately
##--------------------------------------------------------------------------------
# Load example data
tracesRaw <- examplePeptideTraces

## Calculate the SibPepCorr of every peptide
tracesRawSpc <- calculateSibPepCorr(traces = tracesRaw,
                                    plot = TRUE)

## Filter the raw traces object to a protein FDR of 1%
tracesFiltered <- filterBySibPepCorr(traces = tracesRawSpc,
                                     fdr_cutoff = NULL,
                                     fdr_type = "protein",
                                     FFT = 1,
                                     absolute_spcCutoff = 0.2,
                                     rm_decoys = FALSE,
                                     plot = TRUE,
                                     PDF = FALSE,
                                     CSV = FALSE)
summary(tracesRaw)
summary(tracesFiltered)

CCprofiler/CCprofiler documentation built on May 19, 2021, 7:40 p.m.