computational-metabolomics / metfrag-galaxy

Metfrag for Galaxy

MetFrag for Galaxy

Galaxy tool wrapper for MetFrag.

Website: http://c-ruttkies.github.io/MetFrag

Source code: https://github.com/c-ruttkies/MetFrag

Version

• MetFrag tool - v2.4.5+galaxy3
• MetFrag Vis tool - v2.4.5+galaxy0

(Using MetFrag v2.4.5)

Galaxy

Galaxy is an open, web-based platform for data intensive biomedical research. Whether on the free public server or your own instance, you can perform, reproduce, and share complete analyses.

Suspect list

The list of suspects is an aggregated list of in silico predicted MS/MS spectra of natural products from the Universal Natural Products Database (http://pkuxxj.pku.edu.cn/UNPD/index.php). The list is an aggregated version of the github repository https://github.com/oolonek/ISDB/tree/master/Data/dbs.

Developers & Contributors

• Jordi Capellades ([email protected]) - Universitat Rovira i Virgili (Tarragona, Spain)
• Julien Saint-Vanne ([email protected]) - ABiMS (France)
• Tom Lawson ([email protected]) - University of Birmingham (UK)
• Kristian Peters ([email protected]) - IPB Halle (Germany)
• Payam Emami ([email protected]) - Uppsala Universitet (Sweden)
• Steffen Neumann ([email protected]) - IPB Halle (Germany)
• Christoph Ruttkies ([email protected]) - IPB Halle (Germany)
• Ralf J. M. Weber ([email protected]) - University of Birmingham (UK)

Changes

2nd July 2020

• NEW TOOL "MetFrag Vis" - 2.4.5+galaxy0 released

  • Thanks Kristian Peters!
  • Enhanced visualisation of MetFrag results

• Metfrag tool - 2.4.5+galaxy3 released

  • Option added to include all fragment peaks in output
  • Additional meta columns added for output (if available) e.g. retention time
  • Provide adducts at tool level as well as MSP level

Version 2.4.5+galaxy2

• Change owner to computational-metabolomics
• Update tests output based on updates of external compound databases

Version 2.4.5+galaxy1

• Update based on IUC guidlines

Version 2.4.5+galaxy0.1.13

• Submitted to test toolshed (8:9a3019c609d9)
• Improved reporting of adducts
• Removed print output of settings that might reveal sensitive details regarding the database
• Added missing inchikey column to metchem output

Version 2.4.5+galaxy0.1.12

• Submitted to test toolshed (7:0b3816a7a14b)
• Allow metchem database to be configured via a config.ini file (allows more flexibility)

Version 2.4.5+galaxy0.1.11

• Submitted to test toolshed (6:ceb9bd68f6bc)
• Update to MetFrag version 2.4.5
• Boolean for default suspect list fixed
• Check for no results added
• Doc updates

Version 2.4.2+galaxy0.1.10

• Submitted to test toolshed (5:c53ed894d736)
• Add default option for suspectlist (so user does not need to upload the GNPS list)
• Changed "Minimum percentage of explain peaks" argument to float
• Added "skip invalid adducts" option
• Updated unit-tests
• Check for empty input file added

Version 2.4.2+galaxy0.1.9

• Submitted to test toolshed (4:eb581a101672)
• Bug fix for neutral mass calculation (Introduced from 0.1.3)
• Additional adduct forms for negative added
• Version system changed to IUC style (TOOL_VERSION+GALAXY_TOOL_VERSION)
• Memory regex check added to stdio - should now repeat at higher memory if low memory error found
• Update unit-tests to use RP022611 MassBank data (Glucose)

Version 0.1.8:

• removed quotes for output columns (previously double quotes " were being added for inchikeys and scores causing problems for downstream processing.

Version 0.1.7:

• Bug Fix for ID tracking
• Bug Fix for skipping final MSP spectra if two empty lines not present

Version 0.1.6:

• Bug Fix for when NoExplPeaks is zero

Version 0.1.5:

• Fix to add MetChem command line param
• Added auto select for MSP schema
• Added unit test for suspect list

Version 0.1.4:

• added UNPD InCHIkey database to be used for automated testing
• acknowledge additional contributors

Version 0.1.3:

• merge with the latest PhenoMeNal develop version of the module, based on https://github.com/korseby/container-msnbase
• merge with changes of Julien Saint-Vanne

License

Released under the GNU General Public License v3.0 (see LICENSE file)

metfrag-galaxy's Issues

Update based on the msp-to-metfrag PhenoMenal tool

After some discussion with @korseby we are going to update the galaxy tool here (https://github.com/computational-metabolomics/metfrag-galaxy) with the functionality that has been developed as part of PhenoMenal (script and xml). Keeping the tool as python though.

The plan will be that we will later migrate this tool to a central Metfrag repository

@jsaintvanne: What is the status with your pull requests? Can you update or create a new pull request that merges them with the branch "phen_msp_to_metrag_update".

FYI: @RJMW

let metfrag detect out of memory condition

When submitting with to less memory I get:

Error: A JNI error has occurred, please check your installation and try again
Exception in thread "main" java.lang.UnsatisfiedLinkError: /gpfs1/data/galaxy_server/galaxy/database/dependencies/_conda/envs/[email protected]/lib/libnio.so: /gpfs1/data/galaxy_server/galaxy/database/dependencies/_conda/envs/[email protected]/lib/libnio.so: failed to map segment from shared object: Cannot allocate memory

in stderr.

Maybe detect Cannot allocate memory as out of memory error, then Galaxy can react on this.

Default weights used for scoring

Hi @korseby,

There were some default weights given in the Galaxy tool development for the scoring approaches used for FragmenterScore, OfflineMetFusionScore and SuspectListScore.

These were originally given the weights of 0.4,0.6 and 1.0.

These were added when the Galaxy tool was updated from the MetFrag Galaxy tool PhenoMeNal wrapper - I just wanted to see if you know where these weights came from?

implement local db / metfrag as conditional

Since either a local db or a metfrag db has to be chosen this shoud be realised with a conditional.

Running Metfrag as Container fails

The Metfrag tool runs fine when running with a local executor. However, when I use a Docker executor, Metfrag fails. The reason seems to be that it requires Python which is not included in quay.io/biocontainers/metfrag:2.4.5.

I imagine the fix would be to add a Python requirement in <requirements> and Galaxy would make a "mulled container", at least that's what I hope...

InChi error on usegalaxy.eu

Currently getting the following messages from usegalaxy.eu

Error loading JNI InChI native code.
You may need to compile the native code for your platform.
See http://jni-inchi.sourceforge.net for instructions.

org.openscience.cdk.exception.CDKException: Unable to load native code; net.sf.jnati.NativeCodeException: Error creating directory: /opt/galaxy/.jnati/repo/jniinchi/1.03_1/LINUX-AMD64

Do not get the same issue when running locally, on travis and bham.

IUC compatible

As part of the "Hackathon on Galaxy Tools and Workflows for Metabolomics" at EMBL-EBI - we would like to make this tool compatible with the IUC guidelines

move repo to IPB

As discussed at our MetaRbolomics hackathon, move repo to https://github.com/ipb-halle .

Custom adduct types

It would be great if we could provide a more extended list of adducts to use with the MetFrag Galaxy tool. Currently we support only a few adduct types

In the MetFrag documentation "Further parameters" section in mentions "PrecursorIonMode" should be set for different adducts i.e.

Further Parameters
PrecursorIonMode
The adduct type of the precursor is used to calculate fragment masses. Following adduct types can be set by their appropriate numerical value encoding the following types:

positive (IsPositiveIonMode = True)
1	-	[M+H]+
18	-	[M+NH4]+
23	-	[M+Na]+
39	-	[M+K]+
33	-	[M+CH3OH+H]+
42	-	[M+ACN+H]+
64	-	[M+ACN+Na]+
83	-	[M+2ACN+H]+

negative (IsPositiveIonMode = False)
-1	-	[M-H]-
35	-	[M+Cl]-
45	-	[M+HCOO]-
59	-	[M+CH3COO]-

no adduct (IsPositiveIonMode = True/False)
0	-	[M]+/-

So I am wondering if these adduct types are hardcoded? or would it be safe to presume we could add in any other adducts as long as "PrecursorIonMode" is set to the rounded value of the adduct (e.g. [M+Na] is 23) and the correct neutral mass has been calculated for "NeutralPrecursorMass".

Do you know anything more about this @korseby?

@RJMW