Navigation

Introduction

VEMS is a program for analysis of MS-based proteomics data. The current version can process raw data and perform database dependent searches. VEMS can furthermore analysis iTRAQ, Mass tag, SILAC, 18O labelled samples and do label free quantitation. The Mascot importer now uses dta files directly making it very easy to use Mascot for searching and VEMS for quantitation.

References

Please reference (a) suitable article(s) below if you use VEMS for your research

  1. VEMS: iTRAQ tool for database dependent search, quantitation and result storage. Accepted for publication in Proteomics, Jan 4, 2010.
  2. Matthiesen R. Methods, algorithms and tools in computational MS-based proteomics: a practical point of view. Proteomics. 2007, 7, 2815-32.
  3. Matthiesen R, Sorensen M, Bunkenborg J, Hojrup P, and Jensen ON. VEMS 3.0: Algorithms and computational tools for tandem mass spectrometry based identification of post-translational modifications in proteins. J. Proteome Res 2005, 4, 2338-2347.

Please reference the following book chapters if the SILAC correction methods are used:

  1. Matthiesen R, Carvalho AS. Methods and algorithms for relative quantitative proteomics by mass spectrometry. Methods Mol Biol. 2010;593:187-204.
  2. Gruhler A, Matthiesen R. Methods Mol Biol. 2007;367:139-52. Quantitation with virtual expert mass spectrometrist.

License Agreement

Terms of license:

VEMS is provided “as is”, “with all faults” and without any express or implied warranty. In no event shall the authors or IPATIMUP be held liable for any damages arising out of the use of or inability to use this software, even if its authors or IPATIMUP has been advised of the possibility of such damages. If you do not want to accept the terms of this license, you must not install VEMS. By choosing to install this software you are accepting these terms.

Requirements

Download a VEMS versions:

VEMS 5

VEMS 4

Important requirements:

  • You will need Winrar (http://www.rarsoft.com/) to unzip the downloaded folder.
  • Version above 4.20122009 have been tested with screen resolutions from 1024X768 -> 2000X1900. It is recommend to use a screen resolution >= 1280x800.

VEMS:

Windows XP, 2000, Server 2003, server 2007, Vista and Windows 7 have been tested

The regional and language (Control panel) should be set as:

  • English (United States)
  • If the above requirements are fulfilled then you are ready to run VEMS.exe

Exraw_micromass:

Requires Masslynx and PLGS from Micromass

Exraw_Xcalibur:

Requires Xcalibur software

Download SIR:

SIR 0

The regional and language (Control panel) should be set as: - English (United States)

Installation

Download a VEMS versions:

VEMS 5

VEMS 4

Use Winrar to unpack the folder. The VEMS folder can be located in any directory that you have write permission to.

Issues with PLGS and Masslynx

If you get an error that says:

“Cannot open file “C:\data\”

Then please check carefully the following:

  1. make sure that you have installed PLGS in the default directory if not then you need to specify the path of the PLGS software in Exraw_micromass.
  2. language settings set correctly as “English (United states)”.
  3. the correct version of PLGS is specified in Exraw_micromass.exe.

Please also check * Requirements

Documentation

Manual for VEMS 5: Manual

Manual for VEMS 4: Manual

Frequent asked questions: FAQ

Data

iTRAQ test data from:

VEMS: iTRAQ tool for database dependent search, quantitation and result storage. Accepted for publication in Proteomics, Jan 4, 2010.

Test data

Bugs, improvements and updates

18 April, 2011: Bug fix:

  1. I recently discovered that the modification file for iTRAQ modifications that was distributed with VEMS had an error which is unnoticed when searching with iTRAQ-Lys as fixed modification. The modification mass table is now updated and a default amino acid mass table with MMTS-Cys and standard Lys mass can now be selected together with iTRAQ variable modification on Tyr, Lys and N-terminal. You can test that the two ways of searching iTRAQ data basically gives the same result now.
  2. The conversion program Exraw_Qstar had a bug that that caused the program to stop processing, in some cases, when many VEMS raw files was converted to pkx. This bug is now fixed.
  3. Note the latest two VEMS versions require that the searched sequence databases get indexed again. Just run “File -> Convert FASTA for fast MSMS search” again overriding the old indexed database. The old version can still use the old format but the new versions can only use the new format.

28 Mar, 2011: Bug fix:

  1. When the new search engine was inserted into VEMS, in the update from 22, Feb 2011, it was reseting the size of modification array causing an error when trying to update the modification settings after a database search. This is now fixed.

22 Feb, 2011: Updates:

  1. Search speed is now up to 10 times faster.
  2. Lower memory usages when searching large data sets.
  3. Protein inference algorithms are now directly incorporated into VEMS.
  4. The Mascot parser is now more robust and imports CSV files.
  5. Repeated search is now supported.
  6. More proteases are now directly defined in the search engine.
  7. All search parameters can be specified in the startup window with no need to navigate to any other windows.

10 Nov, 2010: Updates:

  1. VEMS now have direct support for 6-plex TMT tag and iTRAQ using MMTS modification on cysteine. The user can now select these mass tables directly in the program.
  2. VEMS now have a isoform resolver.
  3. More filtering options have been included in FDR menu.

20 Sept, 2010: Updates:

  1. VEMS now supports 8-plex iTRAQ (A big thanks to those of you who provided test data).
  2. VEMS now supports customized score functions. The user can define which ion series the program should consider for scoring.
  3. Spectra processing in Exraw_Xcalibur is now improved and tested with more data.
  4. Exraw_Qstar can now process raw data as both profile and centroid to pkx files.

Tutorial describing the new features

1 Sept, 2010: Updates and bug fixes:

  1. An extra search parameter was introduced under: “search settings” which is “percentage of intensity that must be explained”.
  2. A flexible way to calculate FDR thresholds by decoy approach where the random or permutated database is concatenated is now available. It works by specifying a string tag which must be present in the header of all random or permutated sequences.
  3. The FDR score threshold filtering also works on the quantitative results which means that it is no longer needed to repeat the quantitation if new thresholds are used. However, the data that is filtered is lost unless the result is saved before filtering. It is therefore recommended to save the full search result with the quantitative result before doing the FDR threshold filtering.
  4. Exraw_qstar program can now convert mzXML from “Trapper I” program. The mzXML does not contain charge state assignments which are done by Exraw_qstar. Exraw_qstar assumes that the parent ion is given in the mzXML is correct and parses them directly to the VEMS pkx files. It is possible that Exraw_qstar needs to be improved with respect to charge and parent ion determination.

27 July, 2010: Updates and bug fixes:

  1. The false discovery rate is now even lower compared to previous VEMS versions.
  2. The search setting: “Filter isoforms” has now been renamed to “Filter isoforms (Require minimum one unique peptide)”. This setting corresponds to the Mascot setting “Require bold red”. In previous version the filtering was always active but it can now be deactivated.
  3. The minimum length of required b- or y-ion series can now be adjusted. The parameter can be set lower if the mass accuracy on the parent is high.
  4. The navigation in large search results has now been improved and works very fast. A small tutorial is demonstrating how to navigate in results can be found here. An annoying bug, that occurs when the setting “printing single protein hit” is activated, has been eliminated.
  5. Debug information can now be saved to disc by activating “Write debug information to disc”. This setting will make the search slower and should therefore only be used if a VEMS search breaks.
  6. The user settings on peptide and protein score threshold have been eliminated and the user now only have to worry about the mass accuracy settings and the minimum length of continuous y- or b-ion series required.

02 July, 2010: Updates:

  1. All scoring weights applied have been optimized using more reference data sets. You should see significantly more proteins identified at a given FDR threshold compared to what is obtained with the previous version (Using FDR threshold estimated by randomized, reversed or scrambled sequences).
  2. A new function is included that calculates the fraction of explained spectra “Analysis -> Validation -> Percentage of spectra assigned”.

22 June, 2010: Updates:

This update is faster and more efficient in assigning spectra than previous VEMS 5 versions. I have observed up to 1.8 times more assigned spectra to peptides of a single protein using VEMS 5 compared to what was obtained with Mascot. I am so confident with VEMS version 5.22062010 that I will reward the first 5 who points out better search result, when using Mascot instead of VEMS, with a free copy of the book “Bioinformatics methods in clinical research”. To get the book please send me data and search result (only CID data from Orbitrap and Q-tof data will be considered). Please use exactly the same data for comparison. If you need to convert msm files or mgf files to pkx files then you can use the VEMS program mgf2pkx.exe located in the VEMS folder “Exraw”. I will only use the received data for optimizing VEMS and the data will be kept confidential.

  1. Many users do not consider setting the Gaussian weight parameter so that it fits the accuracy of the data (including myself). VEMS now does this automatically for the user. The Gaussian parameters for the fragment ions and the parent ions are now estimated by VEMS automatically. This has a profound effect when searching Orbitrap data where the mass accuracy on the fragment ions is around 0.5 Da. This new automated estimation by VEMS means that you can obtain up to 1.8 times more spectra annotated by using VEMS rather than Mascot. The update will also have a minor effect on the quality of the search result when searching Q-tof data. It is minor since the default Gaussian parameters in previous versions were optimized for Q-tof data.
  2. VEMS uses now less memory. The previous VEMS versions were tested with large Orbitrap data set and searched against databases with sequences from a single species. However, this version has optimized memory usages and have been tested with large data sets searched against large databases with sequences from many species. The memory optimization was done by discarding random matches as soon as the corresponding protein score for the random matches are calculated. The full peptide match information is necessary for the protein score calculation but we are not interested in the specific random assignments in the final report. Therefore the only information on the random assignment left in the end of the search is the peptide and protein score distributions.
  3. VEMS now by default only print assignments to a single selected protein to screen. This is because it takes too long to print the full report to the screen if the data set is large. The full report can still be printed by (Setting -> Tab-page search -> Uncheck “Print only one protein hit to the output window”)

28 May, 2010 Updates:

  1. A peptide modification filter is added to the “Output” window. This filter can be used to view for example only all the assignments with a specific modification or exclude all peptides with a specific modification from the search result.
  2. Henrik Molina and Francesco Mancuso from CRG Barcelona, Spain requested to get scan numbers included in pkx files. Xraw_Xcalibur is now updated to provide scan numbers in addition to retention time in the pkx files.
  3. Eva Rodriguez Suarez CIC bioGUNE, Bilbao, Spain reminded me of the usefulness of getting the score distributions saved so that the FDR threshold no longer needs to be set just after the search. In this VEMS release score distributions are saved as well with the rest of the search result. When the result is loaded the score distributions are loaded automatically as well.
  4. A new scrollbar is added so the user can select the number of proteins and based on the selected number of proteins the FDR and score threshold is calculated.
  5. The function “View -> FASTA handler” is now available. This tool can be used to make sub-specie databases from large databases. For example if you have the FASTA sequence of all Swiss Prot proteins then you can easily make a sub-database with only Human proteins. Some old VEMS function for automatic download of sequence databases will soon be available as well.
  6. A hardware memory problem when performing label free quantitation on large Orbitrap data is fixed.
  7. A problem with FDR calculation when using ppm for mass accuracy has been fixed.
  8. A problem with viewing alternative solution for phospho-peptides in the spectra viewer has been fixed.

14 May, 2010: Updates

  1. A problem with the scoring of the random peptides is now fixed. Furthermore, VEMS has now two numerical methods for calculating the FDR score threshold. The new method is named “Robust estimate of protein FDR” which will be discussed in an upcoming manuscript.
  2. A length restriction on the output is now applied all peptides shorter than or equal to 3 is not shown in the output report.
  3. The program now outputs “Quantitation was successful” if the quantitation completed successfully.

9 May, 2010: Updates based on comments received on VEMS workshop held on BMB, SDU, Odense, Denmark on the 6th of May

  1. Exraw_micromass has been updated. The program now has an auto detect button to detect which version of PLGS is installed.
  2. Exraw_micromass can now be started from any path on the computer. However, if you are using windows Vista then make sure the program have the rights to write on the hard disc.
  3. The main menu has been cleaned so only relevant functions are available.
  4. When closing and starting Exraw_micromass remembers the previous path used for PLGS.

4 May, 2010: Update and bug fix:

  1. After introducing the multivariate regression a bug allowed too many alternative solutions to enter the search reports. This bug affected search results from large data sets (more than 50000 spectra) by running out of memory or degrade the final result. This bug has now been eliminated and several large data sets have been tested again. However, please let me know if you encounter problems with large data sets. In addition, manual validation tools for quantitation works directly on the new uniform data structure. This means that it is possible to jump directly from the “multiple experiment viewer” to the raw data. The previous manual validation tools for quantitation are still maintained.

22 April, 2010: Update and bug fix:

  1. A bug was discovered and sub sequentially removed (the bug prevented the completion of the search) by searching a data set provided by Kerman Aloria, Miren Josu and Jesus M. Arizmendi from UPV, Bilbao, Spain.

15 April, 2010: Update and bug fix:

  1. An exception that was occurring in some cases when searching large databases with high redundancy has been eliminated (result of the error was no reported search result when actually many spectra could be assigned). The redundancy caused the multivariate regression to fail.
  2. Default setting for spectrum plot with overlay of theoretical ions gives now in general nicer default plots.
  3. Xraw_Xcalibur raw data processing is now improved by activating the setting “Make alternatives” which means that in case there is doubt about the parent ion mass and charge then VEMS inserts alternative spectra processing solutions. This gives bigger files but in VEMS 5 this is not really a problem.
  4. Xraw_Xcalibur a bug occurring if both peak selection and deconvolution raw data is selected has been eliminated.
  5. Assigned spectra can now be saved directly to file by right clicking on the assigned peptide in the search report.

29 Mar, 2010: Update:

  1. VEMS 5 is now available. VEMS 5 is optimised for searching large data sets. More than a million spectra can be submitted in a single submission.

20 Dec, 2009: Updates:**

  1. User requested update included. VEMS now automatically adds modifications specified for the quantitation to the modifications for the database dependent search.
  2. User requested update included. VEMS now offers an additional result display showing only non redundant peptides and omitting protein sequences.
  3. Using PPM for mass accuracy gives now search times compareable with using Da

19 Dec, 2009: Bug fix:

  1. VEMS was in some cases hanging when searching large databases with many modifications. This bug is now eliminated.

11 Dec, 2009: Updates:

  1. This version has been tested with screen resolutions from 1024X768 -> 2000X1900. It is recommend to use a screen resolution >= 1280x800.
  2. A minor bug when using the SILAC correction factor for incomplete isotope labels has been eliminated.

18 Nov, 2009: Updates:

  1. Box plots of ratios has been added to the multiple result viewer
  2. It is now possible to define your own experimental designs for quantitative analysis
  3. Data analysis of iTRAQ, label free and 3-plex SILAC use unified data structures and downstream data analysis tools
  4. Sub-areas have splitters so specific areas can be enlarged
  5. A number of extra checks on user input data have been added. The file structures of projects and database files are examined before opening.
  6. Common settings for quantitation is now unified. The peptide score threshold setting works for all quantitative methods and can be set to correspond to a specific peptide FDR threshold.
  7. Label free quantitation is now very easy - just specify your data and press the “quantify” button - if no stable isotope labelled amino acids are specified then VEMS assumes label free quantitation.
  8. All the quantitative methods now removes proteins with no quantified peptides in the quantitative output.
  9. SILAC correction methods now available again for the new data structures

9 Sep, 2009: Updates in responds to Proteomics reviewers comments:

The potato peroxidase test data sets that come with VEMS download can be search in 6 seconds on a laptop with one CPU against all species in SwissProt. For larger data sets the search time is typically 30 seconds to 2 min. The user interface has been cleaned up and all functionality is now documented.

  1. P-value estimation (Mascot and PEAKS Q does not estimate P-values)
  2. The quantitative values from VEMS have in all cases we have tested been more accurate than values obtained by Mascot and PEAKS Q (I am very interested in case where this is not true).
  3. Can summarize several large search results (both identification and quantitation)
  4. Between reporter ion normalization (quantile normalization) and median normalization
  5. Repeated search: It is recommended to use the repeated search functionality by using unspecific cleavage. This will automatically repeat the search and find extra peptides which is beneficial for the quantitation (gives more accurate quantitation).
  6. MySQL storage is now better documented
  7. Sequence databases are stored in 2/3 of the original space of the FASTA file (VEMS does not use the original FASTA file after conversion)

11 May, 2009:

Exraw_Xcalibur is now improved (Thanks, to Henrik Molina, CRG, Barcelona for pinpointing this shortcoming in Exraw_Xcalibur). Exraw_Xcalibur now also checks the mass of the observed parent ion in the LC-MS run (in addition to the charge state determination). The new Exraw_Xcalibur now leads to approx. 15 extra spectra assignments per standard BSA run compared to the old version.

11 May, 2009:

VEMS is now updated with a new function (Analysis -> Quantification -> Improve score if HCD and CID confirm the same hit. Insert peptide for HCD spectrum if only CID spectrum is assigned).

18 May, 2009:

VEMS is now updated with an improved user interface for label free quantitation. There are now two set of functions for label free quantitation. The two sets perform the same analysis and give the same result. However, the new set of functions (marked with low memory tag) can be used for very large set of LC-MS runs since it loads the LC-MS runs dynamically. An updated tutorial is also made available for label free quantitation. Furthermore, spectra processing is now faster (Try to compare generation of single ion chromatograms using an older VEMS version versus this release).

Support

Bugs, questions, comments and suggestions can be addressed to Rune Matthiesen rmatthiesen@ipatimup.pt

Table Of Contents

This Page

Navigation

© Copyright 2010, Rune Matthiesen. Created using Sphinx 1.0.3.