Differences between genomics readme and automatically generated tsv

[somathias]

There are currently the following differences between the information in the genomics readme (which is based on interactive submissions to ENA using the Webin portal) and the automatically generated .tsv (and json) that are based on the information pulled from ENA's xsd files.

• name vs alias: different labels for the same field?
• VIB's scripts pull a dedicated title field in addition to name/alias. Needed?
• file_format vs file_type: different labels for the same field?
  • Moreover, the CV pulled from the xsd files (see eg in ENA_experiment_metadata_fields.json) is much more extensive. Maybe the additional fields are outdated in practice?
• the fields are listed in a different order
• different description texts

How to harmonize?

[YvonneKallberg]

It is the automatically generated .tsv that can be regarded as sustainable, since it will be possible to keep the metadata up to date in the long run. However, thinking from end user / researcher point of view, it might be important that names and descriptions are the same as if the user would do an interactive submission without the help of our templates.

[YvonneKallberg]

name is not the same as alias:

• The name will be shown when data is public, while the alias is a unique identifier that can be used internally.
• The alias is e.g. used when doing programmatic submission (where experiment and run metadata are located in two .xml files), in order to link a specific run with a specific experiment.
• An alias is thus not needed when doing interactive submission via browser (using tsv files) (but I do think that ENA creates an alias for you, when you submit).

However, I'm almost 100% that name and title is the same (without checking description text of the browser versus .xsd sources).

[somathias]

name is not the same as alias:

• The name will be shown when data is public, while the alias is a unique identifier that can be used internally.
• The alias is e.g. used when doing programmatic submission (where experiment and run metadata are located in two .xml files), in order to link a specific run with a specific experiment.
• An alias is thus not needed when doing interactive submission via browser (using tsv files) (but I do think that ENA creates an alias for you, when you submit).

However, I'm almost 100% that name and title is the same (without checking description text of the browser versus .xsd sources).

thanks for the clarification @YvonneKallberg. Does this mean that the tsv we generate should include the name/title field instead of the alias field? If yes, do we use name or title?

title has the description "Short text that can be used to call out experiment records in searches or in displays. This element is technically optional but should be used for all new records." from here: https://github.com/enasequence/webin-xml/blob/8e764aab17b0557a786d72eb18cbe07a4121142f/src/main/resources/uk/ac/ebi/ena/sra/schema/SRA.experiment.xsd#L719

[YvonneKallberg]

I would say yes on inclusion of name/title, and would call it title for 2 reasons

1. when I look at downloaded tsv templates from ENA they don't have this column (I think I got the name field (both the actual field as well as the name 'name') from documentation and it has stuck with me throughout submissions)
2. 'title' is what we get automatically which make life easier.

[YvonneKallberg]

Another difference I've noticed is study and sample fields:

• study_alias from automated but study from interactive (the latter can be populated with either an accession number or an alias)
• sample_alias from automated but sample from interactive (the latter can be populated with either an accession number or an alias)