Additional columns/examples questions

[timosachsenberg]

This is refereing to the protein tables (ae, de) maybe the examples are outdated.

• how to handle protein groups (currently, accession were always one protein)
• consider adding gene columns (my bet is that this will be requested a lot by external users)
• columns between de and ae differ a lot (e.g., ae has more columns with condition, sample etc.). Is this intentional?

[timosachsenberg]

In https://github.com/bigbio/quantms.io/blob/dev/docs/feature.rst

• Especially for TMT data the representation seems to be very redundant. E.g., it looks like for TMT10plex one would have 10 features. Each with all meta data or even PSM information / spectrum intensities.
• mz_array and intensity_array could be renamed to mzs and intensities (or even fragment_mzs, fragment_intensities)

[jpfeuffer]

Regarding duplication. I raised the same concern but consensus was that parquet will only store unique string or at least compress them anyway.
I think this should be benchmarked for the most common use cases.