Entropy panel unavailable if mutations aren't translated

[jameshadfield]

Current Behaviour

For the entropy panel to be displayed in auspice we have three requirements:

1. JSON.panels includes "entropy"
2. JSON.meta.genome_annotations exists and includes at least a {nuc: {start: INT, end: INT}} object.
3. Some mutations to be annotated on branches in the tree.

These typically come from an augur workflow with steps:

(i) augur ancestral (does not create a nodeDataJSON.annotations object)
(ii) augur translate (creates a nodeDataJSON.annotations object)
(iii) augur export (takes care of requirement 1 unless you opt-out, and converts the nodeDataJSON.annotations object to JSON.meta.genome_annotations).

However if you choose not to translate mutations (step ii) then no annotations object is available for export, and thus requirement 2 is not met and the entropy panel is not displayed.

Expected behavior

A pipeline using steps (i, iii) should be valid. In other words, translating mutations should be optional.

Possible solution

The solution is not as simple as just adding an annotations block to the node-data JSON produced by augur ancestral, as augur export v2 assumes that there will only be one of these blocks.

The most consistent would be for

• augur ancestral creates a nodeDataJSON.annotations object with nuc information
• augur translate creates a nodeDataJSON.annotations object with per-gene information
• augur export v2 accepts multiple annotations blocks and combines them, accepting identical duplicate elements and exiting if there are conflicts.

[corneliusroemer]

I got extremely confused by this bug. I encountered it as the following error in a workflow that uses augur translate and just has a .gff as input, not a .gb with nuc annotation.

The error is:

Validating schema of 'auspice/monkeypox_global.json'...
        ERROR: 'nuc' is a required property. Trace: properties - meta - properties - genome_annotations - required
Validation of 'auspice/monkeypox_global.json' failed.

------------------------
Validation of auspice/monkeypox_global.json failed. Please check this in a local instance of `auspice`, as it is not expected to display correctly. 
------------------------

Thrown by augur export. I couldn't figure out how to get it to just read nuc from the .gff, which I assumed was possible. Couldn't believe it isn't. But that's the way it seems to be 😬

This is a very serious bug, it really makes using the same genemap for Nextclade and Nextclade reference build of monkeypox impossible which is not good, discrepancies arise like this.

[corneliusroemer]

Unless the problem described by me is not a real bug, or not identical, feel free to readjust pain score. But based on my use case in addition to the original from @jameshadfield I've upgraded this Crash, blocking and some users. It's crashing my workflow, there's no sane workaround and it cost me a lot of time. I also can imagine this happening to others if they don't use a .gb as input for translate but .gff which is theoretically supported.

[corneliusroemer]

The workaround for me is to:

1. edit the line

   augur/augur/utils.py

   Line 150 in 5916362

   limit_info = dict( gff_type = ['gene'] )

   to

limit_info = dict( gff_type = ['gene', 'source', 'nuc'] )

2. Use the following nuc annotation in the genemap.gff:

MT903344.1	Genbank	source	1	197233	.	+	.	locus_tag=nuc

I'll put in a PR, hope that doesn't break anything. Should I open this as a separate issue @jameshadfield ?

[jameshadfield]

#961 will close the original bug identified in this PR, but it won't address the gff + augur translate bug - let's use #953 to track that one.