I was wondering if it is possible to provide links from the bars of a histogram (to slices of data) that appear on the report page? This is possible for a list widget where you get the option to view results or create list.

Would be really nice if region searches could be fully integrated with queries/templates/lists.

• Use case: want to find candidate genes in the chromosomal regions defined by QTL. Right now user can easily query for the QTL and get their coordinate ranges. But they cannot feed those ranges directly into the region search. (The user would have to reformat the results from the first query in order to paste the ranges into the region search box.)
• Use case: want to keep a list of regions that I can use to drive the region search.
• Use case: want to incorporate region search into a query (analogous to how Lookup is integrated?)

For given genome coordinates, export sequence.
Or export sequences of features with extensions/flankings set by a user (add a input box in results table export popup)

Can GoConverter be modified to store column 16 information i.e. annotationExtension. Potentially also be able to load the Ontologies (cell type ontology) that this column refers to and do all the right things.

http://wiki.geneontology.org/index.php/Annotation_Extension

The docs on intermine.org are in a trac wiki which is partially broken now. Move to github:

1. reorganise
2. identify missing bits
3. docs "closer" to code, more likely to keep updated
4. run a script to update markup

Joel is creating protein coding genes for mouse but genes for human and rat. the keyword search is creating a organism facet for the most specific term (protein coding) instead of the most inclusive (gene or even sequence feature).

On the Genomic Regions Search Results page user may find it useful to be able to pick multiple types or the whole list when exporting. This is a feature request.

I have to keep adding the same column of interest (FBgn identifiers) every time I open a particular list. Would it be possible to save my display preferences for individual lists when I'm logged in?

We can embed list analysis widgets but not report widgets (formerly known as report displayers).

Placing the cursor in the column header makes the column width jump around - especially noticeable for long path names.

May be improved by hiding the pathname with a place holder icon and click to show/ hide.

If I have a gene (or a list of genes) I would like to see genes that are similar to my gene - eg. have a similar expression pattern, go annotation, pathways, etc.

Also, given multiple lists, show me where they intersect through a given element eg. Pathways. This is similar, in principle, to the InterMod tool which does this across species. Here the lists may come from one (or more) species.

A simple feature request to add the asymmetric difference as another option for actions of a list view.

The Regions Tab is not highlighted after clicking on it. All other tabs are highlighted when clicked on; discovered with FireFox.

Can we please have an option to export GO annotations in GAF format?

Source: Flymine

Problem: In-line table text truncated - element style 'max-width' too small.

To reproduce:
Gene report page - D. mel Cks30A

Interactions --> Toggle table. All table text truncated at 2nd character.

<a class="im-cell-link" href="http://www.flymine.org/release-34.0/report.do?id=103034350" data-original-title="" style="max-width: 9.285714285714285px; ">

        cdc2
    </a>

Current model assumes (wrongly) that there is a single unique "withText" for a given GO annotation and that with Objects are pipe separated that need to be added to "with" collection.

This is probably the way to model it (check 2.0 GAF format guide)

• GOID can have one or more evidence codes that can come from one or more publications.
• Each evidence code can have one more "with" information.
• evidence code + "with" is specific to that publication.

example data:

gene_association.goa_human:UniProtKB Q13426 XRCC4 GO:0005515 PMID:16275660 IPI UniProtKB:O00499 F DNA repair protein XRCC4 XRCC4_HUMAN|XRCC4 protein taxon:9606 20120723 IntAct
gene_association.goa_human:UniProtKB Q13426 XRCC4 GO:0005515 PMID:16439205 IPI UniProtKB:Q7Z2E3 F DNA repair protein XRCC4 XRCC4_HUMAN|XRCC4 protein taxon:9606 20060203 UniProtKB
gene_association.goa_human:UniProtKB Q13426 XRCC4 GO:0005515 PMID:16439205 IPI UniProtKB:Q9H9Q4 F DNA repair protein XRCC4 XRCC4_HUMAN|XRCC4 protein taxon:9606 20120723 IntAct
gene_association.goa_human:UniProtKB Q13426 XRCC4 GO:0005515 PMID:16439205 IPI UniProtKB:Q9H9Q4 F DNA repair protein XRCC4 XRCC4_HUMAN|XRCC4 protein taxon:9606 20061120 UniProtKB
gene_association.goa_human:UniProtKB Q13426 XRCC4 GO:0005515 PMID:17396150 IPI UniProtKB:Q8IW19 F DNA repair protein XRCC4 XRCC4_HUMAN|XRCC4 protein taxon:9606 20090814 UniProtKB
gene_association.goa_human:UniProtKB Q13426 XRCC4 GO:0005515 PMID:9809069 IPI UniProtKB:P49917 F DNA repair protein XRCC4 XRCC4_HUMAN|XRCC4 protein taxon:9606 20070820 UniProtKB

The goal is to make interaction data:

1. easy to query
2. easy to understand results
   • reduce duplication of results

Solution:

Moved all duplicate data to "details" collection, eg. Gene.interactions.experiment moved to Gene.interactions.details.experiment. Now gene1 and gene2 pair are unique.

See http://preview.flymine.org for an example.

Further work:

Moving to "details" has removed duplication but our other problems are not solved.

09/08/12,11:49:44 ERROR org.apache.struts.action.RequestProcessor - Invalid path was requested /galaxyExportOptions

I did not change any galaxy related settings. I see the above error in logs and 'Oops page not found' error when I export to galaxy. Not sure when it stopped working.

Same thing in 1.0 beta as well.

Currently users can create and save lists in their own personal user account. It would be nice if users could share their lists with others.

A simple javascript query builder would be nice, something similar to accordb (accordb-intermine.rhcloud.com) for a single data set. Maybe pathways? With other data sets to follow.

The current querybuilder allows the querying of all tables in the database. this querybuilder would only query some fields in some data sets, and only a little more complex than a template query form.

The difficult part would be that this query builder should have the option to query across all of the MODs' intermines.

1. start with a single data set, eg. pathways
2. query across all mines (optionally)
3. looks like accordb
4. put on crossmodel.org
5. simple interface, like template query forms
6. use results tables
   • results / creating lists should link back to the mines though
7. eventually this will be part of the intermine webapp but not any time soon, very beta
8. limited functionality (compared to current querybuilder)

If you have a template displaying on a report page (as a new-style result table), clicking any of the column controls on that table causes the browser to jump to the top of the page.

To reproduce: On FlyMine, go the the gene report page for adh. Scroll down to "Gene --> Orthologues + GO terms of these orthologues." and open the table display. Then sort or toggle a column; when you click, the display jumps to the top of the page.

Source: Flymine preview

Special character (�) - present in source data - breaks results page filtering.

Problem:
Flymine 'Interactions Experiment '
Externally curated molecular interaction network in Drosophila melanogaster.

contains field 'contact-comment' with non-ASCII characters
CNRS, Equipe Bioinformatique des r�seaux de r�gulation, Institut de Biologie du D�veloppement de Marseille-Luminy, 13288 Marseille Cedex 9 (http://www.ibdm.univ-mrs.fr/)

To reproduce:

<query name="" model="genomic" view="Gene.symbol Gene.interactions.details.confidenceText Gene.interactions.details.type Gene.interactions.details.role2 Gene.interactions.details.role1 Gene.interactions.details.confidence Gene.interactions.details.name Gene.interactions.details.experiment.name Gene.interactions.details.experiment.description Gene.interactions.details.experiment.comments.type Gene.interactions.details.experiment.comments.description" longDescription="" sortOrder="Gene.symbol asc">
  <constraint path="Gene.symbol" op="=" value="cdc2"/>
</query>

Column summary on 'Gene.interactions.details.experiment.comments.description' and apply filter on 'CNRS, Equipe Bioinformatique des... ' etc.

No rows are displayed but filtering on other values works fine.

Add option queries / templates from file from query upload form.

In the Manage Columns dialog, when you browse for and add a column to the results, you have to click "Choose" and then click "Apply". I persistently forget to click "Choose" (despite the repeated pain it causes) Is it necessary? Could we omit that button? Thanks!

Go to the report page for a chromosome. From the on-page table of located features, click "Show all in a table". You get an "oops" error message. (Verified locally and at flymine.)

Here's the error log message:

Caused by: java.lang.IllegalAccessException: Couldn't get field "[interface org.intermine.model.bio.Location].primaryIdentifier" (available fields are [dataSets, end, feature, id, locatedOn, start, strand])
at org.intermine.util.TypeUtil.getFieldValue(TypeUtil.java:105)
at org.intermine.model.bio.LocationShadow.getFieldValue(LocationShadow.java:89)
at org.intermine.webservice.server.core.TableCell.getField(TableCell.java:37)
... 32 more
Caused by: java.lang.NullPointerException
at org.intermine.util.TypeUtil.getFieldValue(TypeUtil.java:96)
... 34 more

Currently the reference population in widgets is genes in the entire database. It would be nice if you could change this to be a list.

1. run summarise-objectstore postprocess
2. release webapp
3. log in as superuser
4. navigate to my mine > saved templates

All of the "summarise" links next to each postprocess should be inactive / not-clickable.

This is working correctly in FlyMine but not modMine.

1. Are the templates actually summarised but just the links are wrong?
2. If not, did the "do-summarise" ant target (run as part of release-webapp) not run?
3. If the do-summarise target ran, why did it fail?
4. Did the summarise-objectstore postprocess fail for some reason?

when you click on a widget and click on "create list", nothing happens.

Also, when you click on download at the top, nothing happens. download works correctly when you click on the individual items however; creating a list never works.

See:

1. https://github.com/radekstepan/intermine-lists-client#features for overall tasks
2. https://github.com/radekstepan/intermine-lists-client/issues?sort=comments&state=open for open tasks

Regarding the 1.0 "Download Pop-Up" when viewing the Query Results page. Currently the download options display all formats (tab, xml, gff3, bed, fastsa …), but given the context of the query results data some of the displayed formats don't make sense (ie, if there is no sequence data, then selecting the fasta format produces meaningless results).

Also, a way to configure the format options, so that "fasta" isn't displayed (we're not planning on providing sequence data anytime soon)?

In the query builder you can query for attributes being NULL but not references or collections.

I understand the code for this is ready to be tested but we need docs.

We have configured our SequenceFeature class with the field config

fieldconfig fieldExpr="sequenceOntologyTerm.name" label="Type"

This causes a javascript error* when GeneFlankingRegions are displayed in a table using the 'show all in a table' link from the OverlappingFeaturesDisplayer. The resulting table is empty or missing rows.

• From FireBug:
  error imjs.js (line 943)
  (an empty string) imjs.js (line 945)

1. When you type into an autocomplete field, the browser's default dropdown (show what you've typed into that field previously) appears over the top of the mine's dropdown.
2. The display is a bit off where the dropdown intersects the edge of the box.

Rachel test

http://intrac.flymine.org/ticket/2983#comment:1

It looks like the new result table display is applying an extra round of character escaping in some circumstances.

Here's the scenario. Mouse allele nomenclature uses angle brackets. These strings are encoded at load time, so that "Cftr<tm1Bay>" is stored as "Cftr&lt;tm1Bay>". For the most part, these display fine, I.e., showing the left angle bracket. For example, on the Cftr gene report page, its table of alleles looks fine. Also, on the allele report page, index search results, etc.

However there is one case where (it looks like) the strings are escaped again, so that the left angle brackets display as "&lt;". This is when I use the query builder. If I return allele symbols, they all come out like this: "Cftr&lt;tm1Bay>"

Specifically, mitominer wants to hide Gene.ontologyAnnotations because this collection is a duplicate of Gene.goAnnotations. This confuses their users.

This didn't work:

  <class className="org.intermine.model.bio.SequenceFeature">
    <fields>
      <fieldconfig fieldExpr="primaryIdentifier" />
      <fieldconfig fieldExpr="organism.name" name="Organism"/>
      <fieldconfig fieldExpr="ontologyAnnotations" hideInQueryBuilder="true" />
    </fields>
  </class>

In the QueryBuilder, it would be very handy if the model browser (left hand) tree would expand to the corresponding point when a node is clicked in the query (right hand) tree. I realize the whole QueryBuilder is up for redesign, but I do think this simple enhancement would be extremely helpful in many circumstances.

The popups that show additional information about an object sometimes extend beyond the visible area. Eg, the popup for the bottom row extends below the visible area. Sometimes (but not always) they extend off the right or left side.

Also, the text within the popup often extends out of the visible area. Long paths exacerbate this.

Is it possible to lay those 5 little icons (remove,filter,etc..) horizontally instead of vertically? The Results table header would be compact and nicer.

From a result table, click on "Manage Columns".
Browse for a column and select it.
Click Choose but do NOT click Apply.
Click Browse for column again.
The button freezes.

You can close/repen the dialog to get it working again.

From an enrichment or chromosome distribution widget, one can view the results in a specific category (GO term or chromosome). The result table has some default formatting, and in particular omits theme.css.

Would it be possible to have autocomplete be available on keyword search? Is this something that can be configured or does it need to be implemented. If so, would it be possible. We think it is a useful feature to have.

Currently you can only have one user be the superuser but there should be a way to have more than one.

The following query (return basic gene info for (lookup) Fgf4 where species is one of mouse or human) does not display. The javascript console shows lots of 500 server errors. Same thing happens at FlyMine.

<query name="" model="genomic" view="SequenceFeature.symbol SequenceFeature.name 
SequenceFeature.primaryIdentifier SequenceFeature.sequenceOntologyTerm.name 
SequenceFeature.locations.locatedOn.primaryIdentifier SequenceFeature.locations.start SequenceFeature.locations.end 
SequenceFeature.locations.strand SequenceFeature.organism.shortName" longDescription="Return genome features that 
match a specified string. (ID, symbol, name, synonym) for the specified species." sortOrder="SequenceFeature.symbol asc" 
constraintLogic="A and B"><constraint path="SequenceFeature.organism.shortName" code="B" op="ONE OF"><value>H. 
sapiens</value><value>M. musculus</value></constraint><constraint path="SequenceFeature" code="A" op="LOOKUP" 
value="Fgf4"/></query>

Joel would like to have a scatterplot.

x-axis: human genome y-axis: mouse genome

dots would be where there is a mouse-human orthologue pair

From Results Table --> get code --> Javascript. The file that is save called code looks to be empty.

Few issues:

1. It looks like the current GoConverter only stores publications where a pipe "I" appears i.e. only such references are stored.

SGD_REF:S000073372|PMID:6759872

If there is just a PMID I think it won't be stored. for e.g. in PomBase file..

PomBase SPCC1884.02 nic1 GO:0005887 PMID:10748059

2. Non PMID references are not stored for e.g.

ZFIN:ZDB-PUB-020724-1
FB:FBrf0159398

So, for GOMine I want to be able to store non PMID references and PMID references. It looks like sometimes there is a database reference for PMID (mod files), and a given PMID might have different databasereference IDs in different GAF files. for example PMID x refers to MGI:Ref1 and SGD:Ref2

I am trying to model like so...(xml below is not showing up...)

<class name="GOEvidence" is-interface="true">
  <reference name="code" referenced-type="GOEvidenceCode"/>
  <collection name="publications" referenced-type="Publication"/>  
  <collection name="dbreferences" referenced-type="DatabaseReference"/> 
  <collection name="pubdbrefs" referenced-type="PublicationDatabaseReference"/> 
</class>
 <class name="PublicationDatabaseReference" is-interface="true">
  <reference name="pubRefId" referenced-type="Publication"/>
  <reference name="dbRefId" referenced-type="DatabaseReference"/>
</class> 
<class name="DatabaseReference" is-interface="true">
  <attribute name="databaseReferenceId" type="java.lang.String"/>
</class> 

This requires modifying GoConverter (storeEvidence() method) to add the right collection and not just "publications"...

if (!publicationEvidence.isEmpty()) {
goevidence.setCollection("publications", publicationEvidence);
}

Skyped Julie about this, making a ticket.