---

User Manual for LWBMatch

A profile-based and feature-based hybrid algorithm for retention time alignment of SWATH MS data

Update: Sep. 15, 2016 Authors: Long Wu, Sabine Amon, Henry Lam

If you have any question about this manual, feel free to ask lwuai @ connect [dot] ust [dot] hk for help.

---

Three tools are provided in this package

1. Retention time alignment tool: lwbmatch
2. Generte ground truth from identification results: groundtruth
3. Recall and precision calculator: recallprecision

Build / Compile from source code

Requirments

• Boost (Version >= 1.52)
• CMake (Version >= 3.2)
• g++ (Version >= 4.8.4)

Command lines for build:

$ cd /path/to/sourcecode
$ cmake .; make install

Then we get lwbmatch, groundtruth and recallprecision in /path/to/source/bin.

---

Data Analysis: Run test data set step by step

1. Download and unzip dataset

$ wget http://nesvilab.org/tsouc/DIA_Umpire_SampleDataUPS.zip
$ unzip DIA_Umpire_SampleDataUPS.zip

2. Run featurefinder

2.1 convert each mzXML to mzML

$ FileConverter -threads 18 -in_type mzXML -in  [inputmzXML]  -out [outputname].tmp.mzML -out_type mzML
$ FileFilter -threads 18 -in_type mzML -in [outputname].tmp.mzML -out [outputname].mzML -out_type mzML

2.2 Extract featureXML

$ FeatureFinderCentroided -threads 18 -in [outputname].mzML -out [outputname].featureXML

3. Run LWBMatch

3.1 Create input featureXMLList

FileName: test.featureXMLList Write the two lines into this file, replace the /path/to/outputname with your own filename

/path/to/outputname1.featureXML
/path/to/outputname2.featureXML

3.2 Run LWBMatch with DTW

$ ./lwbmatch -l /path/to/test.featureXMLList -w 2

Attention: To run LWBMatch successfully, both the mzXML and featureXML of the two replicates are needed. LWBMatch need to mzXML to calculate do the first phase alignment by DTW on MS1 spectra.

3.3 Run LWBMatch with LOWESS (old version)

$ ./lwbmatch -l /path/to/test.featureXML -w 1

4. Create ground truth and calculate recall and precision

4.1 Run DIA-Umpire, get [inputfile_Q1.mgf]

$ java -jar -Xmx16G /path/to/DIA_Umpire_SE.jar inputmzXML  [diaumpire_configurefile]

There will be three mzXML with Q1 ~ Q3 in name, we use one with Q1.

Instructions for downloading DIA-Umpire and its default parameter [diaumpire_configurefile]

• Find diaumpire_configurefile from this page
  Fileanme: diaumpire.se_params

• Find DiA-Umpire from this page
  We use: DIA-Umpire_v1_284.zip

• Unzip DIA-Umpire_v1_284.zip and use DIA_Umpire_SE.jar

4.2 Convert mgf to mzXML, [inputfile_Q1.mgf] -> [inputfile_Q1.mzXML] (Proteowizard required)

$ msconvert --mzXML -o [outputpath] [inputfile_Q1.mgf]

4.3 Run Comet, search [inputfile1_Q1.mzXML] and [inputfile2_Q1.mzXML] together

$ /path/to/comet -P[parameterfile] [inputfile1_Q1.mzXML] [inputfile2_Q1.mzXML]

• Find Comet from this page

  We use comet version 2013.01 rev. 0

4.4 Filter by xinteract , to get *.ipro.pep.xml (TPP required)

input: search result file *.pep.xml output: *.ipro.pep.xml For more information, please refer to this page

4.5 Create ground truth from [.ipro.pep.xml]

$ /path/to/groundtruth [filename:2] [mzXMLListtxtName_contains_two_mzXML_filename]  [ipropepxml]  [threshold_for_filter_default_0.9]  [groundtruth_filename]  [dummy_parameter:1]

4.6 Calculate recall and precision

$ /path/to/recallprecision [groundtruth_filename]  [RTAlignResult_with_extension_resu]  [mz_tolerance:0.1] [rt_tolerance: 1 ~ 25]

---