Please wrap the XCMS refineChromPeaks tool about tools-metabolomics

[eschen42]

@lecorguille I can imagine at least two ways to approach this:

• Add code to xcms_xcmsSet.r along with options in the wrapper to set the parameters
• Create a separate tool

I can start working on the first and give you a pull request. You can let me know whether you would prefer to pursue another option or would like to delegate this to someone else.

[eschen42]

@lecorguille I have been reading the XCMS reference manual and have learned that refineChromPeaks can be used for three things:

• Remove chromatographic peaks with too large rt width
• Remove chromatographic peaks based on intensity
• Merge neighboring and overlapping chromatographic peaks

This seriously calls into question my suggestion of integrating it into xcmsSet, so now I am under the impression that creating a separate tool is the better approach.

What do you think?

[lecorguille]

Hi Art,

It's indeed a balance between modularity and complexity.
Is this tool between, findChromPeaks and groupChromPeaks?

Ping @workflow4metabolomics/ms , do you have an idea?

[eschen42]

@lecorguille Yes, the tool is between findChromPeaks and groupChromPeaks

I have gotten this to work in my fork of tools-metabolomics, wrapped as a separate tool xcms_refine at

eschen42@2f3f29d

It passes planemo test . and runs correctly in planemo serve .; I still need to stress test it a bit more before sending a pull request to solicit feedback formally. But, hey, it's working! The drawbacks are:

• It is dependent on bioconductor-xcms 3.12.0 because refineChromPeaks is not found in 3.6.1
• It generates warnings because R 4.0.2 runs but bioconductor-xcms 3.12.0 was built with R 4.0.3
• It does not appear in the workflow graphic ... perhaps it ought not to?

So, we will need to figure out how to negotiate around this obstacle. It strongly argues against integrating with (and de-stabilizing) the findChromPeaks wrapper.

Thank you for any guidance or suggestions that you can provide.

[yguitton]

Hi Never tried that function, but would be nice to add some of the new xcms functions in the pipeline. Thanks Art for your work. Best Yann Le mar. 9 févr. 2021 09:29, Gildas Le Corguillé <[email protected]> a écrit :

…

Hi Art, It's indeed a balance between modularity and complexity. Is this tool between, findChromPeaks and groupChromPeaks? Ping @workflow4metabolomics/ms <https://github.com/orgs/workflow4metabolomics/teams/ms> , do you have an idea? — You are receiving this because you are on a team that was mentioned. Reply to this email directly, view it on GitHub <#193 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABI76KJ7PPGWI667LUAF733S6DW5RANCNFSM4XDVRF3A> .

[melpetera]

Hi @eschen42

Thank you for the detailed feedback!
I have not tested this function yet, but based on your detailed description, I agree with your suggestion to consider it as a dedicated tool. May I have understood correctly, the function can take as input an object generated by a first findChromPeaks run, and "refine" it to obtain as output a similar object, "filtered" according to the parameters passed to the refineChromPeaks-merge function. Right?

Mélanie

[eschen42]

@melpetera You have exactly understood. For further info, please see the three functionalities of refineChromPeaks documented in the "manual" (the first is the one that I need at the moment, but it seemed sensible to wrap all three):

• https://rdrr.io/bioc/xcms/man/refineChromPeaks-merge.html
• https://rdrr.io/bioc/xcms/man/refineChromPeaks-clean.html
• https://rdrr.io/bioc/xcms/man/refineChromPeaks-filter-intensity.html

[eschen42]

@lecorguille @melpetera
I found a bit of a "showstopper": because refineChromPeaks was not available in XCMS 3.6.1, I wrapped the (proposed) xcms_refine tool with a dependency on XCMS 3.12.0. Unfortunately, the .RData file that this produces has an incompatibility with XCMS 3.6.1, so that xcms_merge threw an error when trying to consume the .RData file.

So, it seems that this is blocked until other tools in the suite run under XCMS 3.12.0

[lecorguille]

Sure, we have to bump the xcms version for all the wrappers.
Can you submit a PR ?

And indeed, unfortunately, we will have to update the different test_data RData :/

[eschen42]

@lecorguille Okay, if I understand you, I can submit a PR for this tool.

Here is the question that I have is what to do about the dependency on XCMS 3.12.0, since ordinarily this is handled through the macros. Should I hard code it into the PR and then you fix it during the merge after the switch to XCMS 3.12.0 ?
To put it another way, are we tracking the needed work and our dialog in a PR ?

Fortunately, it accepts the 3.6.1 test data as input because Johannes Reiner made refineChromPeaks backward compatible.

[eschen42]

@lecorguille @melpetera I am awaiting clarification on the content for a PR. Thank you.

@lecorguille Okay, if I understand you, I can submit a PR for this tool.

Here is the question that I have is what to do about the dependency on XCMS 3.12.0, since ordinarily this is handled through the macros. Should I hard code it into the PR and then you fix it during the merge after the switch to XCMS 3.12.0 ?
To put it another way, are we tracking the needed work and our dialog in a PR ?

Fortunately, it accepts the 3.6.1 test data as input because Johannes Reiner made refineChromPeaks backward compatible.

[lecorguille]

Sorry, I though that I had already answer. It seems that I never press the green button.

Yes, we should just starting by bumping the xcms version in the macro.
My guess that the new CI workflow will raise a LOT of R linting errors :(
But after that, we will be able to have an idea about the planemo test themself.

[lecorguille]

Let see what will happen here: #194

[lecorguille]

So the xcms 3.12.0 is ready here and just waiting for review:
#194

You will be able to base your tools on it

[eschen42]

@lecorguille Good news! Thank you very much! I will try to keep an eye on when #194 gets closed.