importMetaPhlAn

[Daenarys8]

The MetaPhlAn importer currently adds "counts," which could be easily confused with real counts or raw read data.
To improve clarity, I suggest renaming this output to "metaphlan" assay (or a similar label) to distinguish it from actual raw counts.

[TuomasBorman]

Thanks, metaphlan outputs relative abundances so that could reflect the name. There is already a parameter to control the name, the default value could be changed (i.e. the name can be "relabundance" as we already use that naming somewhere)

[antagomir]

The difference is that transformAssay(tse, method="relabundance") generates values in [0,1] whereas metaphlan data is often in [0,100]. It is also common that the values do not sum to 100 exactly, even if they get close to it. So this is somewhat different from what users typically expect from "relabundance" in standard (mia) pipelines.

The default name in the metaphlan importer could reflect that, it could be for instance "metaphlan". The examples could then explain and show how to transform that into [0,1] relative abundances if needed.

[TuomasBorman]

"metaphlan" is also fine, as @Daenarys8 stated the most important thing is that people don't get confused

[antagomir]

Yes, "metaphlan" is pretty descriptive. Or it could be "metaphlan_relab". But perhaps just "metaphlan".

Could you implement this @Daenarys8 ?