WormBase-Caltech Weekly Calls August 2014

August 7, 2014

Topic Curation

• Working out pipeline for curation
• Collecting model/pathway diagrams from papers and reviews: should we make ?Picture objects for these?
  • Do we want to display these published diagrams on ?Paper pages and/or ?Topic pages
  • We need to determine copyrights/accessibility to images
• Would be good to automate identification of Review articles (or "non-primary" in general) in the Topic OA
  • This could be done at the bulk import step
  • Would require a new (possibly read-only) field in the Topic OA to indicate "primary" or "non-primary"
• Chris will look into whether "non-primary" articles are automatically excluded from the Curation Status Form
• The "Curation Status Omit" toggle in the Topic OA may become obsolete once we have a field/column for primary/non-primary status
• Current topic : Wnt signaling
• We have a list of 282 papers from PubMed search using "Wnt" search term (not Mesh term) and C. elegans (Mesh term)
• Finding all papers for a category/topic remains an ad hoc approach; different topics are harder or easier to find all papers for
• We can create a Textpresso category for topics, like "Wnt signaling"

WormBase Ontology Browser

• Needs documenting
• Once everything is clarified, will be pushed to the live site

August 14, 2014

WormBase Ontology Browser (WOBr)

• Should be ready but may need some additional testing before pushing to staging site
• Juancarlos will push to staging during meeting
• Curators should test this afternoon (on staging) and report any issues before 5pm

DataBase call

• This morning (and every Thursday that doesn't have a site-wide call)
• Thomas Down testing Datomic (costs money) and Neo4J (very slow)
• May look again at DynamoDB
• No DB has been officially chosen
• Will have an Amazon AWS for collaboration.
• Expand smallace into bigger mediumace.
• Demo off one database for advisory board meeting.

SAB

• Paul Sternberg back next week; will settle travel plans then
• What advisors are attending?
• What feedback do we want to get from advisors?
• How to show curation/DB progress? What stats/numbers to show?
• We want big picture feedback from biologist advisors; what's useful to the community? What should we prioritize?
• Karen: Perhaps a question to ask is "what are the main questions they are trying to answer when they go to the website?". When they explore a gene or protein function, what is it that they would want to see, and how? I don't think we are missing information so much as a lack of integration of the information at the model level, for example, variation phenotype affects linked to altered protein domain function

• Belated apologies from Mary Ann - clash of evening events.

August 21, 2014

Generating a Site-Map for WormBase

• Use a crawler to generate? Output would need to be made human-readable
• We could use the legacy site as a site map

Citace upload report modifications

• wikipage here: Citace_upload_report
• Goals of this report
  • Summary of the uploaded data classes/objects - this summary should be blind to requests from curators and should alert Caltech to missing data classes or severe changes in numbers of objects within preexisting data classes
  • Summary of curation work - this summary should be curator driven, in some cases the summary will require a more involved aceperl query to get at the actual annotation rather than a straight data class object number.
• Can we get a comparison for those data that are curated through postgres? It would be very helpful to be able to compare the changes in postgres with the changes in the Citace upload.
• Can we automate the generation of this report to make it easier to change and track?
• Regardless of manual or automated report generation when there is a model change or data class addition/subtraction, the responsible curator needs to inform Wen of the need for compensatory modification in the report.

August 28, 2014

Curation Statistics for SAB

• Curators should send Chris all stats for their respective data types: total papers curated, total backlog, false positives
• Can we ignore SVM results for certain data types?
• Can we include Textpresso search results for relevant data types?
• Curators that use a Textpresso pipeline: Karen, Ranjana, Xiaodong, (Daniela?), Mary Ann
• Can we get detailed web usage statistics on particular datatypes?
• We want to articulate our priorities to the SAB; get feedback
• RNAi curation could get up to speed in 5 years if we have two FTEs on RNAi curation
• Are there certain genes that have less phenotype coverage that we should prioritize?

Database migration call

• MongoDB, CouchDB, Neo4J, Datomic, OrientDB, postgreSQL, Cassandra candidates
• Neo4J likely ruled out because of slow performance
• Will compare performance of Datomic vs. Postgres and ACEDB, etc.
• Datomic has good history tracking
• Thomas Down has experience with Datomic
• Probably won't go with a relational database
• We should use the Gene page (webpage) as a demo/example of what we want to try to emulate
• Adam (from Lincoln's group) working with OrientDB (graph database)

Citace Upload Report

• Classes/datatypes missing from Citace Upload Report
• Karen started a Wiki page to capture this info: Citace_upload_report
• Curators should take a look and make sure it is properly filled out for their data types
• Columns are present in the table to make requests for certain numbers in Citace Upload Report and/or the Build Report
• Wen will take requests for queries to Citace etc. to add data to report

UniProt linking to and from WormBase

• UniProt would like us to add information about some specific curated datatypes to a file we already supply that maps paper and gene identifiers
• Kimberly and Juancarlos will work on this
• Need to clarify with UniProt how links back to WormBase can/will be implemented

LEGO backend updates

• Still need to get some backend logistics sorted out
• Communication between Michael Muller and Chris Mungall
• OA-like interface for Noctua?