Do we need MZML_INDEXING module

[ypriverol]

Description of feature

This module was external when the mzml is converted outside TRFP; as we are not using mzml statistics anymore, and DIA reads the mzml in their own way, I don't know the impact of creating yet another extra file in the filesystem; the majority of the tools read the mzMLs sequentially and don't need Random access with indexes.

Can we remove this @yueqixuan @jpfeuffer @timosachsenberg

[yueqixuan]

I think that makes a lot of sense. I agree.

[ypriverol]

@vdemichev do you know if indexing basically adding the idex in the mzMLs has some major impact in DIANN. For example imagine someone onverted outside quantmsdiann and starts with a non indexed mzML?

[vdemichev]

I think it's required, but I am not sure. If it manages to load mzML without index, would mean it's irrelevant if index is there.

For example imagine someone onverted outside quantmsdiann

The best practice in general is not to convert at all and supply DIA-NN with vendor-format data.

[ypriverol]

Agree @vdemichev; this will depend a lot on DIANN support in Unix for vendor-format:

• .d (bruker) already supported and no conversion needed.
• In a previous PR Add --mzml_convert parameter for native Thermo .raw support #68 we are making RAW thermo native; no conversion needed.
• Can you let me know if you support .wiff / .wiff.scan for ABSciex in Unix?

Can you double-check with the codebase of DIANN if indexed mzML is needed?

[vdemichev]

Sciex .wiff is Windows-only.

Can you double-check with the codebase of DIANN if indexed mzML is needed?

It's in a dependency library, not our code, but I am pretty sure it's needed

[ypriverol]

mzML with no index; is a valid mzML; if indexing is needed and conversion happens outside for Sciex .wiff because is Windows only; should we removed it @jpfeuffer @yueqixuan and let the user that they need to index anyway; if they need to convert; then they can also index; we can try to use the container that i use for ABSciex to convert to mzML and we do that step and index in the same step removing the need for the file; or we leave this outside the workflow with helpers. I don't see the point of converting outside to then transform inside.