Include pathogen in demix output

[jchorl]

With freyja supporting multiple pathogens, it poses an interesting challenge for use with multiqc. Multiqc ingests the demix outputs: https://github.com/MultiQC/MultiQC/blob/3c956163a70ced059ac60d2af231d96e278bdd96/multiqc/modules/freyja/freyja.py#L30

With multiple pathogens, you may end up with multiple files for the same sample. Today, multiqc can't know which lineages/stats apply to which pathogen. If we just annotate it in the csv output, multiqc would be able to name columns e.g. Top lineages (SARS-CoV-2).

Thanks again for the great tool. I tried to spruce up the test coverage a bit too for demix.

[joshuailevy]

Hi @jchorl - sorry for the delay, was traveling when this came in and totally missed it.

This is a good suggestion, but I worry that it'll break some pipelines that pull in Freyja... we'll have to check some of the helper functions/pipelines that read in Freyja outputs to make sure this won't cause any issues, but then we'll aim to incorporate this! Thanks a ton!

[jchorl]

I worry that it'll break some pipelines that pull in Freyja

Oh no - I'd hope these used the column headers instead of reading fields by position... And I'd hope they handle additional, unknown fields...

I totally appreciate you don't want to break other code. Thanks again for checking.

[joshuailevy]

Thinking more about it, I think it should be fine - very few people are using hardcoded positions, so I think we're good to proceed with this. If people really don't want to get with the times, they can use the older version :)

Thanks again for the suggestions @jchorl!

[jchorl]

Thanks for merging! I'll PR multiqc to handle this once it's released. No rush, it'll be a bit before I can get to that.