preprocess_spectra: Preprocessing spectra for q2e estimation

In ismaRP/MALDIpqi: Estimate Parchment Glutamine Index from MALDI-TOF datasets

Preprocessing spectra for q2e estimation

Description

Performs smoothening, baseline removal and peak detection on MALDI samples. From the peaks, isotopic peaks for a list of peptides are extracted.

Usage

preprocess_spectra(
  indir = NULL,
  metadata = NULL,
  mzml_files = NULL,
  spectrum_name_file = FALSE,
  sps_mzr = NULL,
  make_plots = FALSE,
  peptides_user = NULL,
  smooth_wma_hws = 4,
  smooth_sg_hws = 6,
  iterations = 50,
  halfWindowSize = 20,
  snr = 2,
  k = 0L,
  threshold = 0.33,
  local_bg = FALSE,
  mass_range = 100,
  bg_cutoff = 0.5,
  l_cutoff = 1e-08,
  tolerance = 0.4,
  ppm = 50,
  n_isopeaks = 5,
  min_isopeaks = 4,
  norm_func = NULL,
  q2e = NULL,
  ncores = NULL,
  chunk_size = 40,
  verbose = FALSE
)

Arguments

indir	Folder containing spectra in mzML format.
metadata	Data frame with spectra metadata with at least file column. Ideally metadata has been cleaned before with MALDIzooMS::clean_metadata
mzml_files	Paths to mzML files
sps_mzr	Spectra object
smooth_wma_hws	Half-window size for WeightedMovingAverage smoothing method
smooth_sg_hws	Half-window size for SavitzkyGolay smoothing method
iterations	Iterations parameter for baseline detection.
halfWindowSize	Half-window size parameter for local maximum detection.
snr	Signal-to-noise threshold above which peaks are considered
k	k parameter for MsCoreUtils::refineCentroids()
threshold	threshold parameter for MsCoreUtils::refineCentroids()
local_bg	Whether to further to clean peaks of lists by modelling the local background noise. See MALDIzooMS::peaks_local_bg. Ideally should work with a snr threshold of 0. mass_range, bg_cutoff and l_cutoff only applied if local_bg is TRUE
mass_range	Mass window to both sides of a peak to be considered for backgroun modelling
bg_cutoff	The peaks within the mass range with intensity below the bg_cutoff quantile are considered for background modelling. bg_cutoff=1 keeps all peaks and bg_cutoff=0.5 would only keep the bottom half.
l_cutoff	Likelihood threshold or p-value. Peaks with a probability of being modelled as background noise higher than this are filtered out.
tolerance	Mass tolerance in Da between mono_masses and subsequent isotopic peaks and detected peaks. See MsCoreUtils::closest
ppm	Parts-per-million added to tolerance. See MsCoreUtils::closest
n_isopeaks	Number of isotopic peaks to pick. Default is 5 and the maximum permitted.
min_isopeaks	If less than min_isopeaks consecutive (about 1 Da difference) isotopic peaks are detected, the whole isotopic envelope is discarded. Default is 4
norm_func	Function to normalize the isotopic distribution
q2e	If provided, it adds the theoretical isotopic distribution of peptides with this extent of deamidation
ncores	Number of cores used by the Spectra::MsBackendMzR backend in Spectra::peaksData
spectrum_file_name	If mzml_files are provided, whether to use file names as spectra names. Otherwise, it is assumed the the spectra IDs are in the mzML files' headers.
mono_masses	Array with the peptides monoisotopics masses

Details

Provide the input data either using metadata and indir, or provide paths with mzml_files. You can also provide a Spectra object directly in sps_mzr. If data is provided using more than one of the options, the sps_mzr is used, and then the mzml_files.

The default peptides are the ones from Nair et al. (2022). The paper contains the details on the preprocessing procedure.

Value

A list of dataframes, 1 per sample. Each dataframe has 3 columns, m/z, intensity and signal-to-noise ratio for each of the n_isopeaks from each peptide. Missing peaks are NAs.

References

Nair, B. et al. (2022) ‘Parchment Glutamine Index (PQI): A novel method to estimate glutamine deamidation levels in parchment collagen obtained from low-quality MALDI-TOF data’, bioRxiv. doi:10.1101/2022.03.13.483627.

ismaRP/MALDIpqi documentation built on Feb. 14, 2025, 8:28 a.m.