Column header naming

[peterjc]

I was looking at updating the sample output in the Galaxy wrapper help text, and realised that the v0.0.3 changes went beyond adding new genus columns:

abaizan/kodoja_galaxy@0597e33

Old column headings:

• Species
• Tax_ID
• kraken
• kaiju
• combined

New column names:

• Species
• Species TaxID
• Species sequences
• Species sequences (stringent)
• Genus
• Genus sequences
• Genus sequences (stringent)

The new names are much longer, which can be a usability issue for viewing the output as a table (in Excel or in Galaxy), and yet to me are not any clearer.

For the Galaxy wrapper, we would be able to explain the columns in the tool's help text. For other Kodoja users this can go in the tool's command line help, the README file or other documentation.

Previously you have species level read counts from Kraken, Kaiju, and their intersection. Now there are just two counts for "Species sequences" and "Species sequences (stringent)" which is not as obvious.

What it not clear from the small test cases is what should be present in the genus columns (and if this information is repeated if for example multiple species from that genus are identified). I wonder if two tables is clearer, for species level reporting and genus level reporting?

Also could/should the genus TaxID be included?

[peterjc]

Possible new help text, have not yet checked actual implementation of the counts vs stringent counts:

Updated after reading the code:

abaizan/kodoja_galaxy@8e4e133

[peterjc]

After discussion, the Genus sequences and Genus sequences (stringent) columns ought to be zero in the test data output (rather than blank). These are read counts assigned to the genus but not to ANY species.

Thus if we have 300 reads matching Ipomovirus, of which 45 matched Cassava brown streak virus, and 30 matched Ugandan cassava brown streak virus, both rows of the table would report genus numbers of 300 - 45 - 30 = 225 reads. (I wonder if putting the genus level total might be simpler to understand, here 300?)