Semantic resources project/Meeting notes/2009-09-17

SCF/SWAN collaboration meeting: 9/17/2009

Attendance
Jonathan, Paolo, Elizabeth, Timothy D., Tim C.

Agenda

 * 1) Candidate resources
 * 2) Mouse models
 * 3) SWAN
 * 4) (others under discussion)
 * 5) antibodies
 * 6) proteins
 * 7) public pathways (SBML-formatted pathways from Elizabeth)
 * 8) should create a page for these on the Neurocommons wiki
 * 9) Use Cases
 * 10) Annotations (entities)
 * 11) Amazon
 * 1) Amazon

Elizabeth's Pathway Dataset
A group from a university in Siena received a grant for developing pathway information related to Alzheimers; they put it together using CellDesigner. The goal was to make it freely available: 3 years, and now they need a home for it. Freely available under the CC license, and also to have ongoing curation and updating abilities. Siena Biotech introduced them to Ingenuity; last Thursday, they met with Sean Scott for a description of Ingenuity’s database and system.
 * Is this a useful dataset? (8 files, < 1 Mb of data)
 * JAR: “yeah, we can look at it…. If they’re looking for a home, why aren’t the existing big pathways (hprd and reactome) better places for it?”
 * EW: link to antibodies, publication data?
 * JAR: the question is: how would it be curated going forward?
 * EW: we could recruit eight groups to edit and curate these datasets, using CellDesigner. We don’t know if they do any modeling with it.
 * JAR: worry is that the amount of work to “get something done,” vs. the amount of work to “recast this into ontologies that are up to OBO foundry’s standards,” these are different things.

PC and JAR spend some time worrying about the choice of ontologies, ontologies which evolve over time, and how we choose terms or relations.
 * JAR: refactoring and remodeling isn’t hard – choose a relation and move forward…

A discussion of Use Cases

 * EW: worrying about her understanding of “use case” – is it structured, or is it something more informal?
 * JAR: we’ve kicked around the word, but he understands it as (a) identification of who’s doing something, and (b) identified what they’re going to do with the system, what their experience, and what they’re getting out of it, without (c) any notation of implementation.

Discussion of search logs

 * EW: we have a list of all the search queries that were run.  There are follow-ons that are filters, but we don’t have those. – also the set of things they entered in “desperately seeking antibodies.”
 * Either AB, PC, or TWD can analyze the logs.
 * PC: there are only 20 questions for the “seeking antibodies” entry?
 * JAR: search logs will tell us what to put into a triple store – “we’ll need some kind of way to get from that kind of query to a sparql query.”

More Use Cases

 * JAR: a list of candidate resources on the wiki, and then the question becomes: “which ones can we do?”
 * EW: will we be able to deliver use-cases with the report?
 * Interviewing for Use Cases:
 * TWD: I’ve gotten contacts to two people from SD, and have been promised contact with KR.
 * JAR: should take KR through some pages from SCF sites.

Mouse Models

 * JAR: IEDB – immune epitope database – ten curators, relational database, and they’ve gotten the OBO bug. A lot of their work is done in mouse models.  They’re curating descriptions of experiments, and so they need relations to particular models used in those.  Someone asked “what are the problems with annotating these things.”  The biologists don’t want to say “this is the Jackson lab’s blah-blah,” but they don’t want to give an accession number, because they got the mouse somewhere else and they can’t detect identity.
 * JAR: talk to BP at IEDB: “what did you get from Jackson labs, who was your contact, and what did you do with it?”
 * EW: MS is the technical person at the Jackson Lab

TC has entered the room.

Amazon

 * JAR: we want to make sure a virtuoso instance under the control of this project is ready
 * TC : Marco suggests a different virtuoso instance
 * PC: because they run a different instance of Virtuoso.
 * JAR: just put a different virtuoso on the same machine.
 * JAR: we need to talk about the resources.

TC: a question about the antibody stuff.
 * What would it take to get up some subset that we could test other things with?
 * PC: we have the models, so if we have a source of data we can scrape it and input it. Today we’re looking at the AlzForum requests.

Use Cases Again

 * John W. and TC were both at the socity for neuroinformatic meetings – they agreed to work with Journal of Neuroscience for a demo project on semantic annotation on Neuroscience.  They’re interested in actually annotating *materials*.
 * Assuming that someone has annotated against antibodies – what kind of enriched information would an author like to see?
 * TC: IP and IS (precipitatino and staining) and LIT(erature)  mouse models, antibodies.
 * JAR: we have NIF in the neurocommons
 * PC: parts of OBI / OBO will be relevant to this.
 * TC: we’ll go see Brad.
 * TC: Neurolex/NIF -> mouse models -> antibodies -> genes -> SWAN “this all fits together.”

Action Items

 * EW will forward the pathway file she has to the group.
 * EW will forward the set of search queries.
 * TWD: plan interview for KR.
 * EW: find the link to the “desperately seeking antibodies” question responses.
 * TWD: look at IEDB for re-done version of JAX catalog. Email BP and MS.
 * TWD: Draw up the list of candidate resources on the wiki
 * TWD: go through the virtuoso installation process for Neurocommons.
 * TWD: information on AWS processing and pricing.
 * TWD: arrange to see Dr. J-H C and BH.
 * TWD: OBO ontologies relevant to ChIP-*