old Stacks 1.4** don't output "Full denovo" output collection

[yvanlebras]

Dear IUC,

Now that new Stacks 2.* seems to be ok, I am trying to retest old 1.4 Stacks version (1.46 notably), notably for the GCC2019 training thursday....

It seems that for genetics map usage, denovomap tool don't populate the Full output from denovo_map on ${on_string} output data collection at least on usegalaxy.eu but I think this is not related to the instance.

        <collection name="all_output" type="list" label="Full output from denovo_map on ${on_string}">
            <discover_datasets pattern="(?P&lt;name&gt;.+\.(tags|snps|alleles|matches))\.tsv$" ext="tabular" directory="stacks_outputs" />
            <discover_datasets pattern="(?P&lt;name&gt;.+\.(haplotypes|genotypes|markers|hapstats|sumstats|sumstats_summary))\.tsv$" ext="tabular" directory="stacks_outputs" />
            <discover_datasets pattern="(?P&lt;name&gt;.+\.(genotypes))\.(loc|txt)$" ext="txt" directory="stacks_outputs" />
        </collection>

For the training, as I am testing the 2.** since a quite long time, maybe the best will be to accept Stacks 2.4 PR so I can give the training using it ?

[bernt-matthias]

So, what exactly is missing? The whole collection or just part of it?

[yvanlebras]

collection is empty, as in this history https://usegalaxy.eu/u/ylebras/h/stacks-rad-genetic-maps

[bernt-matthias]

So the problem seems to be that the tool generates a link batch_1.genotypes.tsv -> batch_1.genotypes_1.tsv

https://github.com/galaxyproject/tools-iuc/blob/b395fa36fa826e26085820ba3a9faacaeddcb460/tools/stacks/macros.xml#L125

Discovery of this data set fails them which renders the whole collection empty (because of an exception raised during discovery)

https://github.com/galaxyproject/tools-iuc/blob/b395fa36fa826e26085820ba3a9faacaeddcb460/tools/stacks/stacks_denovomap.xml#L221

bug described here: https://github.com/galaxyproject/galaxy/issues/8259

For the tool we need to know what batch_1.genotypes_1.tsv is and why it is not generated. Possible bug fixes would be

• to remove the link in the tool script (then it would be missing in the collection)
• touch the link and thereby create an empty output
• extend the filters.

Also one could think of restructuring the output since its highly redundant (eg the content of all_output is contained already in the other collections)

[bgruening]

@yvanlebras @bernt-matthias anything I can do here?

[yvanlebras]

Matthias already have had a look to it last week... Duno if you can manage something but one idea is that we create this data collection as a user.... If this works, it's ok I think

[bernt-matthias]

I think the first step to a good solution is to find out if batch_1.genotypes_1.tsv should be there and if yes: why it is missing.