HCI Data Source: markdown, Import:, Section:


MSICorrect is a Java-based commandline tool that performs recalibration of mass spectrometry imaging datasets without the need of specifying a lock-mass value. The computational approach implemented in the tool generates a reference spectrum, termed here as a consensus spectrum, using all the spectra present in the dataset. This consensus spectrum is used as a reference spectrum to recalibrate the input mass spectra.


[1] Purva Kulkarni, Filip Kaftan, Philipp Kynast, Aleš Svatoš and Sebastian Böcker
Correcting mass shifts: A lock-mass-free recalibration procedure for mass spectrometry imaging data.
Anal Bioanal Chem, 405(25):7603-7613, 2015.

[2] Sebastian Böcker and Veli Mäkinen
Combinatorial Approaches for Mass Spectra Recalibration.
IEEE/ACM Trans Comput Biology Bioinform, 5(1):91-100, 2008.

Download link

MSICorrect v1.0.0 can be downloaded as a jar file here.

The complete source code can be found at GitHub.

Running MSICorrect

You can run the jar file using the commandline on any operating system with java installed. To run the tool, use the below command:

java -jar /path/to/MSICorrect/MSICorrect.jar

MSICorrect does not require any input arguments while running the jar file. All the mandatory user input is taken one by one at the command prompt. You can always use the --help option to get a documentation about the necessary user input. To access the help, simply use the below command:

java -jar /path/to/MSICorrect/MSICorrect.jar --help

Below is the information on mandatory user input:

  1. File path for input mass spectra
    Diretory path that contains a list of peaklists as .txt files that have to be recalibrated. The peaklists should have coordinates values x and y separated by tab in the first line and then the mass and intensity value pairs separated by a tab from the next line and onwards.

  2. File path to store recalibrated spectra
    Diretory path that will store the recalibrated peaklists. Make sure that you have permissions to write the files in the specified directory.

  3. Mass threshold for distance calculation between two peaklists
    Suitable mass window (in Da) depending on the dataset, to calculate pairwise similarity between the two peaklists (example: 0.5).

  4. Peaklist ordering approach (use TG or CG)
    Before performing recalibration, it isimportant that all the peaklists are ordered in aspecific manner. This especially is helpful to minimize the error within the growing consensus spectrum. MSICorrect implements two ordering approaches: type TG for topological greedy approach and CG for Crytal growth ordering approach.

  5. Line distance to perform line-pair stabbing recalibration
    The Maximum Line-Pair Stabbing (MLS) algorithm performs outlier detection and recalibration of a pair of peaklists. This parameter is a distance measure that separates the computational geometry interpretation of two parallel lines in geometrical space (example: 0.5).

  6. Mass threshold for for line-pair stabbing recalibration
    Suitable mass window (in Da) depending on the dataset, to perform recalibration of two peaklists (example: 0.8).

Input file type

As an pinput, MSICorrect requires a directory path containing mass spectra. The spectra should be be preprocessed and peak-picked before being submitted for recalibration. Preprocessing and peak-detection of imaging MS data be performed using several packages available in Matlab or R. The peaklists generated should be exported as individual tab-separated text files with coordinate values for each spectra comprising the first line of the first. A short example of such a file is provided below:

34  18
101.7920    112.253425
103.6095    1278.583291
104.7244    145.525605
105.6506    263.697254
106.7781    129.368985
107.7910    105.195891  
108.7951    127.642998
109.9640    157.623051
110.7048    166.225015
111.8753    118.473873
112.8081    115.421191

Sample data

Sample data to run MSICorrect can be found here. To use the sample data, dowload and unzip the folder. The sample data contains 1184 peaklists in text file format. To use the sample dataset, launch MSICorrect using the command provided about. The user input screen should look like the below:

  __  __   _____  _____  _____                               _   
 |  \/  | / ____||_   _|/ ____|                             | |  
 | \  / || (___    | | | |      ___   _ __  _ __  ___   ___ | |_ 
 | |\/| | \___ \   | | | |     / _ \ | '__|| '__|/ _ \ / __|| __|
 | |  | | ____) | _| |_| |____| (_) || |   | |  |  __/| (__ | |_ 
 |_|  |_||_____/ |_____|\_____|\___/ |_|   |_|   \___| \___| \__|

version 1.0.0       Contact: Purva Kulkarni ()

Commandline usage: java -jar MSICorrectV_1_0_0.jar
Access help: java -jar MSICorrectV_1_0_0.jar --help

MSICorrect needs the following user input
File path for input mass spectra: Sample_data/
File path to store recalibrated spectra: Recalibrated_spectra_output_folder/                 
Mass threshold for distance calculation between two peaklists: 0.5
Peaklist ordering approach (use TG or CG): CG
Line distance to perform line-pair stabbing recalibration: 0.5
Mass threshold for line-pair stabbing recalibration: 1.0

MSICorrect will display the following steps while recalibrating the peaklists

STEP 1 of 6: Reading directory path and parsing data files
Total data coordinates: x = 37 y = 32
There are 756 empty peaklists in the provided mass lists.
STEP 2 of 6: Now performing distance calculation.
STEP 3 of 6: Generated pixel scores for all co-ordinate positions. Now performing peaklist ordering.
STEP 4 of 6: Ranked list generated (Size: 429). Now generating final consensus spectrum.
STEP 5 of 6: Final consensus spectrum generated. Now recalibrating the input peaklists against the final consensus spectrum.
STEP 6 of 6: All new recalibrated files successfully generated!

The recalibrated files can be found in the provided output folder.



  • release version