(BJ) Items to remove from the drop down menu for Pombemine

[ValWood]

These will never be annotated for fission yeast of provided by PomBase, so to display them will be confusing for users:

cDNA clone Chromosomal Deletion Chromosomal Duplication Chromosomal inversion Chromosomal Translocation Chromosomal transposition Chromosome based Chromosome structure variation Component (we will not import because it is a small amount of arbitrary /inconsistent data) Oligo EST Primer Reverse primer Golden path fragment PCR product Overlapping EST set Enhancer ( not relevant for yeast, could be taxon restricted) Sequence collection Ortholog evidence code (unless this refers to source, currently it is not populated- we were going to load orthologs from other sources too?)

We might use these in the distant future, but better not to display things for which data is unavailable: Synteny block Syntenic region Transposable element insertion site TF binding site Strain Sequence variant

[ValWood]

It is also not good to display the groupining terms for "Ontologies" Ontology annotation Ontology evidence code etc

It doesn't make sense to do a query to return all as GO terms and phenotypes are distinct (apples and oranges). Also phenotypes are associated with alleles, and GO terms are associated with genes and they have different metadata. So having this as an entry point from a menu doesn't make sense.

[danielabutano]

These will never be annotated for fission yeast of provided by PomBase, so to display them will be confusing for users:

cDNA clone Chromosomal Deletion Chromosomal Duplication Chromosomal inversion Chromosomal Translocation Chromosomal transposition Chromosome based Chromosome structure variation Oligo EST Primer Reverse primer Golden path fragment PCR product Overlapping EST set Enhancer ( not relevant for yeast, could be taxon restricted) Sequence collection Ortholog evidence code (unless this refers to source, currently it is not populated- we were going to load orthologs from other sources too?)

We might use these in the distant future, but better not to display things for which data is unavailable: Synteny block Syntenic region Transposable element insertion site TF binding site Strain Sequence variant Yes @ValWood you are right. BG interface shows all the classes defined in the model, instead to show only the classes that have data (as the old interface does). We need to improve it @uosl @rachellyne do we have a ticket on BG for that?

[heralden]

BG interface shows all the classes defined in the model, instead to show only the classes that have data (as the old interface does). We need to improve it

@danielabutano This has been explored in the past and it doesn't seem to be doable to have it like the old UI querybuilder with the current web services (see last comment in https://github.com/intermine/bluegenes/issues/662).

[ValWood]

That's a shame. I think it is really confusing to have lots of data types visible that are not relevant to the resource. For many of these, it isn't immediately obvious what they are either from the labels so the user needs to click on them to see if there is data. Also some data classes (like evidence code) do not make a useful starting point for queries and so this is also misleading.

Presumably, the drop done menu could be made configurable. In fact, it must already be configurable be to some extent because there are some data types that we intend to load that are not yet visible as classes (modification), and some datatypes that would not be relevant for other species (PECO- this will be removed for PomBase, it's not the correct namespace)

For example near to the top of the list here:

...only 3 data classes will ever be relevant. It will be very difficult for bench biologise users to get to grips with the interface if these are not hidden.

[heralden]

I may have misunderstood. I was thinking of the model browser tree, where the old UI will hide class nodes that don't have any data.

For the root classes which are displayed in the dropdown, it's easy to tell which are populated or not, so those can be hidden. I have created a ticket for this here: https://github.com/intermine/bluegenes/issues/770

[ValWood]

Cool, the model browser tree is not such an issue because here it is obvious what is populated and what isn't. It would be nice if the non-populated data types were slightly greyed out. With the current blue/ black distinction it isn't so immediately obvious, especially at the top of the model:

A user could easily think that only the 'high-level' classes were blue. it takes a bit of browsing to figure out that only populated classes are blue.

[heralden]

I agree it would be better if these were light grey rather than black. Created a ticket for this: https://github.com/intermine/bluegenes/issues/775

[danielabutano]

The new BG version fixed it

[ValWood]

Hi we aren't seeing this fix. Unless I misunderstood what was happening here. We still see all of the unused datatypes looking normal in the menu.

All of us found this really confusing.

Will reopen, and can discuss if required.

@manulera @kimrutherford

[ValWood]

I think we were talking at cross purposes. We were hoping that the classes that are not populated would not appear in this menu:

[ValWood]

Actually, I now see that the completely unpouulated datatypes are now absent from the menu (thanks for that, it is much better)

However, even some data types that might be imported do not make sensible starting points for queries, which leads to lots of futile attempts to build queries that are not possible or meaningful. It would be much better if we could provide a list of sensible query starting points to be configured in the drop-down.

It will make the whole process of building queries from scratch much easier for our users.

I think some items are populated with a very small amount of inconsistent data and so we can probably suppress a few more by tweaking the import. I will look into that first.

[ValWood]

Closing, it is much better. I don't think anything can be done about displaying populated terms.

[heralden]

@ValWood You can select data types to be added to the group directly below Select to start query which only consists of Gene right now. Unfortunately this can only be done from the old webapp: open http://legacy.pombemine.rahtiapp.fi/pombemine/tree.do and login as an admin, then add the tag im:preferredBagType to terms you wish to have in that group. These same terms will also be displayed at the top of the List type dropdown on the upload page.

Daniela created an issue to add this to Bluegenes https://github.com/intermine/bluegenes/issues/699 but sadly it got lost with time.

[ValWood]

Got it. Se we can move all of the starting points that we think people will use up. I will try this.

[ValWood]

I can't see how to log in as Admin in the old interface. I use our admin e-mail, but it it says I am logged in as user "staff" and I don't see any admin options available?

[ValWood]

AT the very bottom of the page it says "Admin mode", but I don't see where to go...

[kimrutherford]

Hi Val. I see lots of "Add tags" buttons. If you don't try going to: http://legacy.pombemine.rahtiapp.fi/pombemine/tree.do

[ValWood]

I do see that. But in that list I do not see any. of the terms that I want to add to the 'top group' (this would be all of the terms we have used to begin queries so far

GO phenotype allele

etc

And I don'ts ee any way to drill down (if they are hidden)

[kimrutherford]

And I don'ts ee any way to drill down (if they are hidden)

They seem to be mosltly underneath "Annotatable":