Semantic resources project/Meeting notes/2010-01-14

In Attendance
KT, EW, JAR (by phone), SD, TC, PC, TD

Updates

 * review - DEFER

TC presentation

 * 1) SWAN-SCF integration - expect to start in February.
 * Main thing needed: SWAN 1.2
 * 1) StemBook Phase II - start date February, but doesn't mean the resources need to be done by then.
 * Get to updated SCF platform, then make more of a community, annotate with community resources, use PRO and antibodies, mouse resources, etc. Annotation is independent of SCF platform.
 * AR - Request: Set of use cases, tech requirements would be useful for antibodies resources, PRO and mouse resources.


 * 1) Neuropathic Pain Forum -
 * Main holdup is hiring a site editor - have someone IDed and expected to have them start in Feb.
 * Use cases pending discussion with new editor.
 * TC expects the use cases for this will be very similar to Stembook Phase II


 * 1) MS Forum -
 * Still a proposal, has gone to funder, should know by Monday, January 18 if final approval.
 * Requires, by the beginning of Q4, a version of SWAN v.1.2


 * 1) SWAN Incorporation into NIF - promised to have data feed for 1.2 by February


 * 1) SCF Semantic Annotation of J Neuroscience
 * work has been done to get licenses for J Neurosci content
 * requires antibody resources, including alzforum, other antibody data, especially exactantigen, mouse resource
 * requirement - ability to add new antibody with a new target
 * TC was thinking ID existing sources of antibody info that seem comprehensive enough to use as a basis for doing annotations for J Neurosci - even if just a hyperlink
 * requirement - resource has to be represented. need to figure about how to do annotation, may be some UI work / process (sees string, finds that representation in the semantic resources)
 * requirements (get strings that rep antibodies that are resolved - comprehesive and coverage) vs implementations (exactantigen content and exactantigen itself)
 * required coverage - when you take string and return a list of antibodies, the correct antibody is there/correct URIs. request from maryann --
 * desire to have it be as comprehensive in terms of coverage as exact-antigen


 * 1) PD Online Research Phase II
 * Not sure what requirements are.
 * May be beneficial if we intersect the requirements, for example, the fact that they are creating a DB of mouse models
 * Working with ClinicalTrials.gov, perhaps review to see if should be considered.


 * 1) Autism Research Forum
 * Feb 4 - will know more


 * 1) HSCI Blood Program
 * Antibodies are used to sort the diff cell types, but is a very different antibody.
 * AR says its within scope since involves cell sorting.
 * SD to post survey of antibodies from HSCI that they are already using, then, take maybe a dozen and match with catalog numbers, based on this antibody what sort of experiments have been done.
 * maybe 50 total


 * 1) Pathways:
 * goal to represent pathways in a way that integrates with PRO, GO, etc
 * AR talking about his existing work with Lindsey Cowell
 * What take advantage of tools and methodology and apply to some of the pathways in the scope of this project -- he's hoping to have some of the OWL out this month

PRO (15-20)
Uniprot identifiers in PRO.
 * Issue of APP proteins:
 * AR: issue of species specific proteins forms was being held up by uncertainty with PRO, already said they were going to support it so they (PRO) need to make good on their promise to add
 * Alan asked PRO for
 * info for uniprot and pro coordination
 * what's the relationship between pro terms and uniprot terms?
 * A Uniprot identifier identifies..."something." It's not stated anywhere.
 * We can surmise that it identifies something.
 * The goal is to figure it out, confirm it with Uniprot, and then use
 * The problem is that Uniprot has canonical sequences, but there is also the sequence clusters.
 * it's not clear how far out the similarity should be extended to figure out class membership.
 * The current structure wrt isoforms is that there are &quot;incomplete&quot; isoform ids.
 * There's a question of whether you can look at uniprot records and figure out whether an isoform contains something with a particular AA sequence.
 * We have a plausible idea as to what a PRO identifier can be. Now what we need to worry about using that in PRO, need to check for more specific, in terms of coordination - want to get sanction from UniProt group (don't want a moving target)
 * Uniprot identifiers seem to pick out classes of gene products from a particular species.
 * What PRO did most recently, go thru UniProt - create across classes orthlog classes
 * UniProt does not have individual specificity, but does have species specificity. PRO would represent the protein forms as opposed to the mutations without the information that links then together
 * A few complications that keep in mind when using UniProt IDs in PRO :
 * 1) Duplicate identifiers merge records, call 1 of the accession the primary, the others the secondary. Which UniProt ID to use? What about the others? Need for coordination.
 * 2) Change management.
 * 3) Protein identity - Some are amino acid - amino acid identical in species
 * 4) How strains are treated? What's the status of a strain?
 * Coordination:
 * UniProt consortium - members from a number of groups. Responsible for putting together proposal to get funded. Cathy is a member of that consortium, brings up PRO and PRO status to them
 * crafting RFP now, funded internationally. Cathy said it was not likely that any progress on substantive issues before this is set in the spring
 * Cathy said main conversation point of late re: UniProt and Pro is on reciprocal linking and coverage
 * Cathy's decision - respond to specific requests, represent and give ID to any proteins that our use cases require - antibodies and APP. She will do that without using UniProt identifiers and still give them a PRO ID and cross references. Will streamline process.
 * Try to get agreement on what the UniPro identifiers identify, focus on coordination
 * Cathy / PRO want to make sure they they can make an automated request, that the form has places for all the info we need to specify and that they can fulfill those requests. AR will discuss next week with them.
 * TC do want SWAN 1.2 to point to UniProt, PRO or both? Have to make sure that everything in SWAN exists in PRO first.

Mouse Models

 * DEFER

Action Items

 * SD - Requirements/ Use Cases for Stembook Phase II (ongoing)
 * SD - Requirements / evaluation for ClinicalTrials (for next meeting)
 * SD/AR - take the diagram of harvard stem cell diagram and extract cell types from that, so they can be added as a semantic resource
 * EW to follow up with Don about progress / status update
 * TD/AR to pull list of proteins listed in SWAN, see which ones have UniProt (PRO as well?) ids