Skip to content

Manual

Jump to bottom
EvgenyZhvansky edited this page Aug 17, 2020 · 24 revisions

  1. Download .mat file data examples from here: https://yadi.sk/d/fswyNmd8taKGmw. You could prepare your own .mat file from .mzXML files as described in the 11 section of this manual. You could convert raw MS files to .mzXML files as described in the 12 section of this manual.

  2. Download all files from https://github.com/EvgenyZhvansky/R_matrix/tree/master/mzXMLReader/bin/Release in any directory and run SSM_display.exe file.

  3. Create a cosine similarity spectra matrix with the SSM_display software by selecting .mat file that was created in a previous step by pressing “Open data” button.

  4. One could select any pixel on the cosine measure matrix to view the corresponding spectra of that pixel. Also one could select another pixel marked with a circle cursor by pressing arrow keys. Titles of figures contain the names of files and the number of a spectrum from this file (after moving median filtration if it was applied). The cosine similarity measure is also reflected above the spectra (in the top-left corner of the right panel with spectra as R=...). The colormap could be changed in Settings by the user choice

The upper graph pane reflects the spectrum in the column, and the lower one reflects the spectrum in the row of SSM.

  1. Colormap could be changed in the Settings window. It is possible to apply different colormaps or to present the matrix as a grayscale image.

  2. One could zoom in graph pane for one spectrum or both spectra by checking the checkbox “Same X scale”. To unzoom spectra, one could use double-click. While zooming the area one could press the “+” or “-” button to include or exclude the current m/z range from the cosine measure calculation. If the “+” button is clicked before excluding anything, it excludes everything except the active range

Unzoom after button “+” click (by double-clicking on a graph pane):

  1. By removing the checked ranges of included and excluded m/z ranges from the list-box, one could modify the existing mesh of the m/z array.

Press “Remove” to delete checked range (so the range is altered from “+” to “-” and vice versa):

Press the “Reset ranges” button to clear all ranges and to calculate the SSM in a whole m/z range.

  1. Threshold level. Threshold level. This slider can be used to filter noise intensities from being taken into account. It can take the values between 0% and 10%. 0% percent means that all intensities are taken into account for SSM calculation. 10% means that in each scan, intensity values lower than 10% of their maximum in the scan will be set to zero.

  2. Outliers level. Outliers slider allows excluding those pixels (columns and rows), whose sum of values is less than the value of the outliers level. Outliers level could be varied to exclude those spectra (both rows and columns of these spectra) from consideration, whose sum of similarity values to other spectra is less than the selected percent of range of this sum for all spectra. Outliers level varies from 0%, when all of the scans will be considered, to 100%, which means that only one scan will be left (of course, that doesn’t make sense, but values under 100% could be considered).

The range of outlier levels is 0 - 100%. 0 corresponds to the minimum of the sum of R values in columns (or rows as they are equal), and 100 corresponds to the maximum.

  1. “Save” button allows saving the cosine measure matrix view as a .png, .jpg, .bmp, or .gif file or both SSM view and two spectra, or stream the cosine measure matrix and other data in .txt file. Also one could call the context menu by clicking the right mouse button on the graph panes with spectra. There are four options to save the results: 1) saving figures of both spectra similarity matrix and two spectra corresponding to a selected pixel in an automated way by “Save images”; 2) saving figures of the spectra similarity matrix with choosing name and extension of the file by “Save image as…”; 3) save two spectra in .emf vector format through the context menu of the graph pane; 4) saving data in a text file in automated mode by “Save data”. Data is tab-separated in the last option and includes coordinates of clicked pixel, filenames of the corresponding measurements and scans, mass spectra of selected pixels (both included and excluded from calculations separately), m/z array and SSM.

  2. Convert one or multiple mzXML files to .mat format via executable file mzXML2mat.exe (https://yadi.sk/d/sGVGPpEjst9Ujw). Requires 1GB of free space in the system partition. First, select the desired mzXML files in the open file dialog, then follow the instructions in the command window and wait until the process is finished. It will create a “data.mat” file in the same directory, where mzXML files were taken from. The file could be renamed in the future by the user. The parameters are listed below. Conversion will be processed after setting the parameters. Wait for the "Press any key to exit... " message.

It will create a “data.mat” file in the same directory, where mzXML files were taken from. The file could be renamed in future by the user. The parameters are listed below. Conversion will be processed after setting the parameters. Wait for the "Press any key to exit... " message.

m/z min: float, optional Lowest m/z cutoff: Mz lower than m/z min would not be converted and processed. The default value is 200. Acceptable values m/z > 0. m/z max: float, optional Highest m/z cutoff: Mz higher than m/z max would not be converted and processed. The default value is 1000. Acceptable values m/z > m/z min. m/z bin: float, optional The bin width of m/z array: a step of m/z array, which will be used for binning spectra to the vector in space from m/z min to m/z max with step m/z bin. The default value is 0.1. Acceptable values m/z < (m/z max - m/z min). median width: int, optional The number of scans to be averaged with a moving median filter (https://doi.org/10.1038/s41598-018-37560-0). Moving median filter width and step are equal to the median width. The default value is None. If the value is None then median width is recalculated in a way to binned median spectra not to exceed allowed space in RAM. Acceptable values from 0 to the minimum number of scans in measurements. GB_max: int, optional The allowed maximum of RAM space (in GB) to be allocated for scans stored in measurements. The default value is 2. Acceptable values from 1 to the actual free RAM space. Notes: If median width is specified, the number of scans to be taken into account from each measurement will be calculated due to RAM limits. So, only the first scans from each mzXML file will be processed for the RAM limit to be not exceeded. For example, if each scan in any measurement would be 1GB size, n_median=1, and each measurement consists of 10 scans (total 2 measurements), then only the first scan of each measurement will be placed in the output dataset due to the 2GB RAM limit. Whereas skipping n_median (n_median=None) would lead to automated n_median calculation, which means n_median would be set to 10 in this example. The total number of scans in output (N) affects the time or even the possibility of the SSM_display to make calculations. Higher N will lead to the size of the calculated SSM as N2. Thus, the user should control the N for not to exceed RAM in the SSM_display process.

The python script mzXML2mat.py could be used otherwise for the same purposes in the same way. The script is developed in Python3.5.2 with packages described in Requirements.txt. The most important function here is read_and_convert_data(filenames, mzmin=200, mzmax=1000, mz_bin=0.1, n_median=None, GB_max=2), which performs calculating m/z array for binned spectra; binned spectra; median filtering of binned spectra.

Parameters: filenames: list A list of mzXML files. mzmin: float, optional Lowest m/z cutoff: Mz lower than mzmin would not be converted and processed. The default value is 200. Acceptable values m/z > 0. mzmax: float, optional Highest m/z cutoff: Mz higher than mzmax would not be converted and processed. The default value is 1000. Acceptable values m/z > mzmin. mz_bin: float, optional The bin width of the m/z array: a step of the m/z array, which will be used for binning spectra to the vector in space from mzmin to mzmax with step mz_bin. The default value is 0.1. Acceptable values m/z <(mzmax - mzmin). n_median: int, optional The number of scans to be averaged with a moving median filter (https://doi.org/10.1038/s41598-018-37560-0). Moving median filter width and step are equal to the n_median. The default value is None. If the value is None then n_median is recalculated in a way to binned median spectra not to exceed allowed space in RAM. Acceptable values from 0 to the minimum number of scans in measurements. GB_max: int, optional The allowed maximum of RAM space (in GB) to be allocated for scans storing in measurements. The default value is 2GB. Acceptable values from 1 to the actual free RAM space. Returns: mz_array, binned_spectra, scans_count: tuple mz_array: m/z bin values. binned_spectra: ndarray of shape [total_number_of_scans, number_of_bins], which corresponds to binned spectra in mz_array. scan_count: ndarray of shape [number_of_files], each number in the array corresponds to the number of scans in each file. Notes: If n_median is specified, the number of scans to be taken into account from each measurement will be calculated due to RAM limits. So, only the first scans from each mzXML file will be processed for the RAM limit to be not exceeded. For example, if each scan in any measurement would be 1GB size, n_median=1, and each measurement consists of 10 scans (total 2 measurements), then only the first scan of each measurement will be placed in the output dataset due to the 2GB RAM limit. Whereas skipping n_median (n_median=None) would lead to automated n_median calculation, which means n_median would be set to 10 in this example. The total number of scans in output (N) affects the time or even the possibility of the SSM_display to make calculations. Higher N will lead to the size of calculated SSM as N2. Thus, the user should control the N for not to exceed RAM in the SSM_display process. 12. Convert any mass spectra files to files in mzXML format via MsConvertGUI from the ProteoWizard project, which is available from here: http://proteowizard.sourceforge.net/download.html It requires Microsoft.NET 4.7.2 Framework to be installed. The program should be used for raw (MS1 level) files to be converted with the parameters that are indicated in the screenshot below:

The parameters could vary due to the version of the MSConverterGUI. The most important parameters to be set are in the Options block.

Clone this wiki locally