error if feature table samples aren't found in metadata

[jairideout]

An error should be raised if any sample IDs in the feature table aren't present in the sample metadata. This is to match the behavior in QIIME 1 and 2 where metadata IDs can be a superset of the table IDs, but every table ID must have corresponding metadata.

forum x-ref

[Oddant1]

Is this still a relevant issue? The error definitely isn't being raised, but it looks like all of the methods that even use metadata in gneiss are deprecated.

[thermokarst]

It might still be relevant:

• https://docs.qiime2.org/2019.7/plugins/available/gneiss/gradient-clustering/

[mortonjt]

Thanks @thermokarst - yes this is still an outstanding issue

It'll boil down to something as follows here: https://github.com/qiime2/q2-gneiss/blob/master/q2_gneiss/cluster/_cluster.py#L93

difference = set(table.index) - set(gradient.index)
if difference and not ignore_missing_samples:
            raise KeyError("There are samples not included in the mapping "
                           "file. Override this error by using the "
                           "`ignore_missing_samples` argument. Offending "
                           "samples: %s"
                           % ', '.join(sorted([str(i) for i in difference])))

Concerning deprecation - the stats methods are being deprecated in favor of other differential abundance methods such as songbird and hopefully aldex2. CC @ggloor

[Oddant1]

And this should be a thing in all q2-gneiss methods taking a Metadata or MetadataColumn as input?

[mortonjt]

Yes - especially if it isn't deprecated

…

On Fri, Aug 2, 2019, 5:18 PM Oddant1 ***@***.***> wrote: And this should be a thing in all q2-gneiss methods taking a Metadata or MetadataColumn as input? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#25?email_source=notifications&email_token=AA75VXJU7AGQLJUR6T6TQ73QCSQCDA5CNFSM4EJ5FLFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3O3RSA#issuecomment-517847240>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA75VXP65KZQF4HEYSMNV3TQCSQCDANCNFSM4EJ5FLFA> .

[Oddant1]

😆 those poor lonely deprecated methods

[Oddant1]

I'm not finding any relevant methods other than gradient_clustering

[Oddant1]

Also, do we want ignore_missing_samples arguments to be included in any relevant methods?

[mortonjt]

Why not - that's the parameter used in emperor. There's definitely the use case that you do want to remove non overlapping samples.

…

On Fri, Aug 2, 2019, 5:39 PM Oddant1 ***@***.***> wrote: Also, do we want ignore_missing_samples arguments to be included in any relevant methods? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#25?email_source=notifications&email_token=AA75VXNXYR77IKHNNUN6AI3QCSSQ7A5CNFSM4EJ5FLFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3O4WUA#issuecomment-517851984>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA75VXMYMUDNJME6PORFQM3QCSSQ7ANCNFSM4EJ5FLFA> .

[mortonjt]

Not quite - that's testing to see if all of the samples are dropped. The code previously tests to see if there is a complete overlap in samples

…

On Fri, Aug 2, 2019, 6:22 PM Oddant1 ***@***.***> wrote: is this exception already doing something similar to this? [image: image] <https://user-images.githubusercontent.com/10642616/62401675-0c8f6c00-b539-11e9-9062-80f11b187b31.png> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#25?email_source=notifications&email_token=AA75VXMBSSCAJDCILHAHMJLQCSXQRA5CNFSM4EJ5FLFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3O6ZWQ#issuecomment-517860570>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA75VXOWHCT5SILJ4L42GPDQCSXQRANCNFSM4EJ5FLFA> .

[Oddant1]

Yeah I realized that after looking at it a bit more closely

[ggloor]

yes we are getting close to having ALDEx2 up and running inside QIIME2. https://github.com/dgiguer is just putting the finishing touches on this