pClean

pClean is a powerful tool to preprocess high-resolution tandem mass spectra prior to database searching, and aimed at filtering out extraneous peaks with/without specific-feature, which integrated three modules, removal of label-associated ions, isotope peak reduction and charge deconvolution, and a graph-based network approach. pClean is supportive to a wide array of instruments with all types of MS data, and incorporative into most data analysis pipelines.

Resources and executive environment

pClean is programed in Java and R, and released as a R package.

Software download: The source code is available at https://github.com/AimeeD90/pClean, and the released software is downloadable at https://github.com/AimeeD90/pClean_release. Please download the latest version.

Java version: 1.8 or later

R version: 3.5.0 or later

Operation platforms: Windows, Mac OSX, Linux

Hardware: 2 CPUs, 4 Gb memory (the more, the better)

How to use pClean

3.1 Data transformation

pClean accepts MGF files as inputs. A vendor-specific format can be easily converted to MGF format using MSconvert of ProteoWizard library.

3.2 Installation

pClean was released as a R package and distributed through GitHub. The installation steps are listed as follows.

• Check R and Java to ensure the required version is installed. If not, you can download the latest version of software from https://cran.r-project.org (or https://www.rstudio.com) and https://java.com/en/download/, respectively.

Note that if you are a Windows user, please add Java path to the system path after the Java installation.

3.2.1 Install the pClean package from a local .tar.gz file:

• Firstly, please download pClean_0.1.0.tar.gz from http://metax.genomics.cn/pClean/files/pClean_0.1.0.tar.gz (or https://github.com/AimeeD90/Public_Packages/archive/master.zip)

• Then, install pClean package using the following command:

install.packages(file_path,repos=NULL,type="source") # please replace the file_path with the real path of pClean package; eg:install.packages("F:/pClean_0.1.0.tar.gz",repos=NULL,type="source")

3.2.2 Install the remote pClean package from GitHub using devtools package:

• Open R software (recommended RStudio), and install package "devtools" via commands:

install.packages("devtools")
library(devtools)

• Install pClean package using the following command:

devtools::install_github("AimeeD90/pClean_release")
library(pClean)

• If you already have a previous version of pClean installed, you can use that to install the development version:

remotes::install_github("AimeeD90/pClean_release")
library(pClean)

• Note that: Sometimes an error ("Http error 403") would be occured when you install the remote pClean package from GitHub using devtools packagethe, it is better to try more than once, or try again another day.

Now pClean is executable on your work station.

3.3 Usage

Here, one fraction of TTE dataset (peptide labeled with iTRAQ8plex) and one fraction of Jurkat dataset (label free) were used as examples to illustrate how to use pClean.

3.3.1 Parameters

All the parameters of pClean are listed in the following table.

Parameter	Description	Default value
mgf	Input MS/MS data	NULL
itol	Fragment ion tolerance	0.05 (Da)
outdir	Output directory	./
mem	The maximum Java heap size, unit G	1
cpu	Allowable number of CPU	0 (all)
aa2	Consider mass gap of two amino acids	TRUE
mionFilter	Filter out immonium ions	FALSE
labelMethod	Peptide labeling method	NULL
repFilter	Filter out reporter ions	FALSE
labelFilter	Filter out label-associated ions	FALSE
low	Clearance of low b-/y-ion free window	FALSE
high	Clearance of high b-/y-ion free window	FALSE
isoReduction	Heavy isotopic ions reduction	FALSE
chargeDeconv	High charge deconvolution	FALSE
largerThanPrecursor	Filter out ions larger than precursor’s mass	FALSE
ionsMerge	Merge two ions of similar mass	FALSE
network	Graph-based network filtration	FALSE
plot	Plot ions-network	FALSE
idres	Identification result, mzid or dat file	NULL
ms2tolfilter	Fragment mass error tolerance filter limit	1.2
debug	keep the edge and peak files for looking up	FALSE

pClean provide with a function to eliminate the immonium ions from MS/MS data, and the list of immonium ions are got from reference. Filter out the immonium ions.

3.3.2 pClean treatment on label-based MS/MS data

1. Open R and load pClean, type:

library(pClean)

2. Set parameters then run pClean:

mgffile<-system.file("extdata/", "tte.frac1.mgf",package="pClean")
pCleanGear(mgf=mgffile,outdir="tte/result",mem=2,cpu=0,mionFilter=TRUE,labelMethod="iTRAQ8plex",repFilter=TRUE,labelFilter=TRUE,low=TRUE,high=TRUE,isoReduction=TRUE,chargeDeconv=TRUE,largerThanPrecursor=TRUE,ionsMerge=TRUE,network=TRUE,debug=FALSE)

3. The resultant MS/MS spectra are written to the ms/ms directory in separate files. To merge all the files, run this:

mergeMGF(dir="tte/result/msms",name="tte.frac1.pClean.mgf")

3.3.3 pClean treatment on label-free MS/MS data

1. Open R and load pClean, and type:

library(pClean)

2. Set parameters then run pClean:

mgffile<-system.file("extdata/", "120426_Jurkat_highLC_Frac1.mgf",package="pClean")
pCleanGear(mgf=mgffile,outdir="jurkat/result",mem=2,cpu=0,mionFilter=TRUE,isoReduction=TRUE,chargeDeconv=TRUE,largerThanPrecursor=TRUE,ionsMerge=TRUE,network=TRUE,debug=FALSE)

3. The resultant MS/MS spectra are written to the ms/ms directory in separate files. To merge all the files, run this:

mergeMGF(dir="jurkat/result/msms",name="Jurkat.frac1.pClean.mgf")

3.3.4 Visualization of ions-network

Optionally, if you want to visualize the construction of ions-network graph, and annotate ions with corresponding peptide fragment, you need do a database search in advance. At present, pClean supports parsing identification results from dat and mzid. The provided sample mzid file was generated using MSGF+ software. To fulfill this purpose, please use the following commands:

mgffile<-system.file("extdata/", "tte.frac1.mgf",package="pClean")
datfile<-system.file("extdata/", "tte.frac1.asc.dat",package="pClean")
pCleanGear(mgf=mgffile,outdir="tte/result",mem=2,cpu=0,mionFilter=TRUE,labelMethod="iTRAQ8plex",repFilter=TRUE,labelFilter=TRUE,low=TRUE,high=TRUE,isoReduction=TRUE,chargeDeconv=TRUE,largerThanPrecursor=TRUE,ionsMerge=TRUE,network=TRUE,plot=TRUE,idres=datfile,debug=FALSE)
mergeMGF(dir="tte/result/msms",name="tte.frac1.pClean.mgf")

mzidfile<-system.file("extdata/", "tte.frac1.mzid",package="pClean")
pCleanGear(mgf=mgffile,outdir="tte/result",mem=2,cpu=0,mionFilter=TRUE,labelMethod="iTRAQ8plex",repFilter=TRUE,labelFilter=TRUE,low=TRUE,high=TRUE,isoReduction=TRUE,chargeDeconv=TRUE,largerThanPrecursor=TRUE,ionsMerge=TRUE,network=TRUE,plot=TRUE,idres=mzidfile,debug=FALSE)
mergeMGF(dir="tte/result/msms",name="tte.frac1.pClean.mgf")

Once the progress completed, pClean creates a png directory and a gml directory. You can match a png or gml file to the corresponding MS/MS spectrum with the help of spectrumInfor.txt (under the directory: tte/result/).

3.3.5 Other filters

Alternatively, pClean implements two reported filters in it, Top10 filter, a traditional intensity-based preprocessing method, and CRF filter (reference), a chemical rules-based approach but unavailable currently.

To use Top10 filter, run the following command:

To use CRF filter, run the following command:

Contact Us

If you have any questions, suggestions or remarks, please contact: aimee.d.ym@gmail.com.