Hi @tmushayahama

Petra from dictyBase is testing the form and would like to be able to autocomplete on her organisms' anatomy ontology, DDANAT.

The DDANAT terms are available in the graph editor, but not currently in the Noctua form.

Jim has recently done some work to bridge the DDANAT ontology with CARO so we should be okay on that front.

Can you check to see why the DDANAT ontology terms are not available in the Noctua form?

Thank you!

From this issue geneontology/go-shapes#52 and today's GO CAM discussion, how does NF integrate all this to the current state?

• Current causal relationship editor
• Creation of new activities
• Reading already created models with has_part
• Extra Operations, Validation and Deletion

tagging @vanaukenk @pgaudet @thomaspd

@tmushayahama
I think we had talked about this at one point, but it would be nice for curators to see the title of their model at the top of the form.

Is this doable?

Thx.

So far there is no way to add base BP and CC after submitting the form.

Without a space after 'DDX58', only the UniProtKB isoform accessions are returned; with a space, the isoform and canonical accessions are returned.

The clone evidence functionality is really handy, but it looks like each use of a given evidence code and reference gets a new entry in the list, resulting in duplicate entries.

Can we make sure the list only contains unique entries?

Currently if a model is not titled, the default naming is "enabled by [GP]"

Coming up with a title on the fly might be hard. Maybe a friendly non-blocking dropdown with few suggestions like below

tagging @vanaukenk @thomaspd you once had a naming suggestion. Although noctua search will solve most problems, visually it is nice to have nice human readable titles

After editing a field in the table, the order changes so that the most recently edited row is moved directly below the MF row.
It would be nice to be able to preserve the same order as what is in the entry forms, i.e. BP always follows MF.

We should remove the 'part of' closure for the autocomplete fields, as it will potentially be confusing and could possibly lead to annotation errors.

Confirmed with @ukemi this morning.

@tmushayahama - let me know if you want to touch base about this at all.

Thx.

See also: #19

Coming from geneontology/noctua-form-legacy#22 with GP

To clarify for curators what entity they are annotating, we will need to also display the unique id or accession associated with the gene name and species.

This will be the case for all other terms

@vanaukenk @thomaspd from last week's discussion. I just realized one problem with autocomplete which was the reason we didn't do this before. Let'say you type lipid transport then it fills in with lipid transport (GO:0006869). However if you wanna change to something like lipid ... then you have to backspace all the (GO:0006869) before you can see results

This will be helpful for larger models so that curators don't need to individually open each activity unit, one at a time.

Form pre-validation or before save checks

• GP is required
• MF is required for Activity Unit and BP Only (root mf)
• if any term is provided, that term evidence is required
• if evidence is provided then reference is required
• DB: Accession check for reference

The BP only form allows curators to select one of six different MF-to-BP relations.

It would be helpful to allow editing of a selected relation in the table by including the same set of relations in a drop-down menu on the table view for a BP only annotation.

Thank you!

@sabrinatoro

Since CC only annotations cannot be causally connected to other annotations, we should remove this functionality from the associated entries in the table:

The idea here is to allow curators to select BP terms like 'cell fate specification', input a specific cell, and then have the correct annotation extension relation selected behind the scenes for use in the model and presentation in the graph editor.

For this we will need to define the exact domain and range constraints for each term, relation, and object of the extension.

@tmushayahama

Would it be possible to auto-size pop-up windows?

For example, if the browser window is below a certain size, curators need to scroll in the Clone evidence box to click on 'Done' after they've selected a reference.

Right now, the form is not constraining how references are added which could result in mis-formatted reference entries.

We had proposed splitting the reference field out into two separate fields:

1. A list of valid dbxefs

2. A blank field for ids/accessions

There would need to be a value for each for SAVE to be enabled.

A bonus would be to retrieve a title for PMIDs to help the curator verify that they've entered the correct id.

Add date to each line of annotation in the form

Add to the right-most part of the form

Add actual date instead of icon

Name of column: Date

In the BP only table view, we need to populate the evidence and reference for the MF row based on the evidence and reference for the BP.

We had this in Noctua form 1.0:

Hi
I wanted to add cerebrospinal fluid UBERON:0001359 and when I typed in cerebrospinal fluid this ID was not available in form but was available when I used the graph editor
Thanks
Ruth

Adding and editing should only be allowed for authenticated users. Although this will not happen, UI should be intuitive .Hiding form, edit links etc

Currently, the BP only version of the form is not allowing me to save an annotation and gives this error message:

even when root MF is automatically entered.

As per our discussions on the workbenches calls, we will remove the second CC from the CC section of the default form but then also add a slot for organism.

Following on from the discussion on the GO-CAM specifications call, we agreed to replace
'located in' with 'part of' for the gp-to-term relation for CC terms.

@tmushayahama We'll need to update this in the CC only version of the form as well as make sure that the relation is then translated correctly in the graph editor.

Thx.

To delineate the autocomplete box better from the form, let's add a background color, e.g. HEX
#FCF3CF?

So curators can check things like the term definition, it would be helpful to have a link out to AmiGO term pages.

To address the issue here:

geneontology/noctua-form-legacy#49

We will, by default, retain any information curators have added but not saved prior to switching forms.

It doesn't look like the 'Remove evidence' option in the table view is working.
When I select it, nothing happens to the evidence associated with the GO term.

To help curators select the correct GO evidence code, it would be helpful to also provide the three-letter GO code in the autocomplete list.

I'm not sure how or where the information is stored for the evidence code autocomplete (i.e. as a dbxref like GOECO:IDA or 'has related synonym' IDA) but it is possible to search on nearly all of the three-letter GO codes and have the autocomplete return the correct code in the list. It would just be nice for curators to be able to visually hone in on the GO three-letter codes more easily.

Also, the only GO code that doesn't present the correct option in the autocomplete list when I type the three-letter code is 'IDA' and I'm not sure why that is. We need to address this, though, as IDA is very commonly used.

After a user has logged in, they need to also be able to log back out :-)

The graph editor displays a logout button next to the user name, and we should add the same functionality to the Noctua form.

Thx.

We need to have the functionality to add a 'NOT' qualifier to annotations.

I think the simplest way to do this would be similar to what we have on the current form in the additional actions list.

We could make a top-level menu of 'Qualifier' and then have 'NOT' be available as the selection choice.

This option would need to be available on every annotation line for the main aspects, Molecular Function, Biological Process, and Cellular Component (not on the additional contextual fields).

Also include 'NOT' for the BP only and CC only forms the same as above, i.e. only the main aspect, not the contextual fields.

@tmushayahama - let me know if you want to discuss further.

There is also a corresponding ticket on the noctua-form-legacy tracker here:
geneontology/noctua-form-legacy#95

Comparing the "search database" option between DDX58 (UniProtKB:O95786) and its isoform (UniProtKB:O95786-1), it turns out that for the isoform no annotation could be retrieved. Is it intended that no annotation is propagated to isoforms?

At the top of the form there is a GO icon and also the text Noctua form.

Clicking on the GO icon should take users to geneontology.org and clicking on Noctua should take users back to the landing page.

Thx.

The current production form allows curators to select evidence to clone from a previous annotation even when that annotation may not have been saved yet.

We will want to have the same functionality on form 2.0.

Thx.

If in a model, an activity unit is NOT connected to anything else, it can be deleted from within the form.

If, however, an activity unit has a connection to other nodes, then the curator will get a warning and NOT be allowed to delete. Information on the exact connection will be provided, wherever possible.

The workflow here would be that they'd then go to the graph editor to update the model.

The 'Copy to Form' function fails to account for skipped lines, such that copying this annotation:

results in this info in form, where the anatomy term 'retina' is now placed in the 'cell type' line (note, I had manually edited the gene product field before taking the screen shot; it copied appropriately)

It saves just fine to save it as is without moving the anatomy term from the cell type line to the anatomy line, but it is a a poor curation experience.

There must be no hard-coded URLs in any workbench, except for invariant external resources. In this case, the hard-coded internal variables mean that the workbench can only be used in production (or some other specific environment).

See: #47 - with this necessary PR, the workbench can only work in production -- if used in dev, it could cause user confusion or data corruption.

This is a critical issue and needs to be prioritized over any other work.

@tmushayahama @vanaukenk

@tmushayahama

Can you double-check the date that is returned when a PMID is entered into the reference field?

I just entered PMID:30789901 and the date was listed as January 31, 2019 although I don't see that date associated with the reference anywhere in the PubMed record:
https://www.ncbi.nlm.nih.gov/pubmed/30789901

Also, it looks like the last author is listed (e.g. Dennis H. Kim) but I thought we said we wanted the first author, et al. (e.g. Fletcher M., et al.).

Thanks!

Hi @tmushayahama

Li from MGI has reported a bug wrt evidence not populating the table when entered into the form, but for a specific case when there is an unpopulated field in the form.

How this arose:
The curator was using the CC only version of the form and entered a gp and a CC. They then accidentally selected the 'Organism' extension when they meant to select 'Cell'.

Since, right now, we don't have a line-by-line delete option in the form (or an undo action), the curator options are to either 'Clear' and start over or just select 'Cell' from the extension menu, leaving 'Organism' in the form but blank.

The curator chose the latter option, filled in the 'Cell' extension and then clicked on add 'Anatomy' and entered an anatomy term and evidence.

Once the annotation was saved, it appeared on the form, but the evidence code and reference for the 'Anatomy' extension was not there. The curator had to add that evidence and reference again from the side 'Add evidence' option.

In trying to replicate this, I've also had the evidence disappear from the 'Cell' but stay with the 'Anatomy' term.

Here's a screenshot of what this looks like in the table:

Note the evidence code and reference missing for the 'oocyte' annotation.

@tmushayahama - let me know if you want to talk and share a screen to troubleshoot.

For example, for BP only -> Create New BP

For CC only -> Create New CC

Add 'occurs in' CC to the BP only version of the form.

The additional contextual information (e.g. a second CC, cell, anatomy, organism) would be selected from the right-hand menu as for the activity unit form.

The link outs to AmiGO from the Noctua form for WB genes gives a 500 error.

The URL currently generated is:

http://amigo.geneontology.org/amigo/term/WB:WBGene00009180

The correct URL is:

http://amigo.geneontology.org/amigo/gene_product/WB:WBGene00009180

Should this URL syntax be in the datasets.yaml file? Or am I missing it?

https://github.com/geneontology/go-site/blob/master/metadata/db-xrefs.yaml

Note that this error arises with other gps as well, e.g. MGI genes.

@tmushayahama

Can we align the allowed values for 'has input' and 'has output' in the BP and MF ShEx shapes with what is allowed in the autocomplete in the form?

For Biological Process:
has_input: ( @<ChemicalEntity> OR @<AnatomicalEntity> OR @<ProteinContainingComplex> ) *;
has_output: ( @<ChemicalEntity> OR @<AnatomicalEntity> OR @<ProteinContainingComplex> ) *;

For Molecular Function:
has_output: ( @<ChemicalEntity> OR @<ProteinContainingComplex> ) *;
has_input: ( @<ChemicalEntity> OR @<ProteinContainingComplex> ) *;

It is possible to save an annotation in the CC only version of the form, if the CC field is not filled in.

We need to implement the same functionality here as in other configs of the form whereby if a required field is not filled in, i.e. GO term, evidence code, reference, the curator cannot save and gets an explanation as to why.

Thx.

From 2019-08-13 call

We looked at the side menu of curator actions for the form. Two main considerations:

1. Available actions are as intuitive as possible to curators
2. We display only those actions that are essential in as neat a way as possible

General summary of considerations from @tmushayahama:

Mostly the option are categorized in the following actions for both the term and evidence:

Add (term or evidence)
Remove, (nested term or any evidence)
Clear Values, (each line, the whole term and its evidence(s) list)

Fill in helpers e.g.
quick fill in values like Root MF
cloning evidence
search database
ND evidence

all other ones which on go rule (later)

Validation (shex, go rule (term, evidence combination) etc)

Information: links to term or evidence page or maybe only on autocomplete

Help : link to documentation

In the Table View of annotations, we'll want to populate the Assigned By field with the Group for which the curator is curating when they make their annotations.

In this example screenshot, the Assigned By field should be populated with WB.

If I am logged into Noctua, when I switch from the graph editor to Form 2.0, my login status is not carried over.

If I try to then login from Form 2.0, I can go through all the usual steps (i.e. login via github) but when I'm returned to the page I'm not actually logged in.

A search in the BP field of the Activity Unit and BP only forms returns some MFs, likely due to 'part of' links between MFs and BPs in the ontology.

We will want to restrict the BP field search solely to BP terms.