Ewan Birney wrote on Wed, 2 Apr 2008 08:28:01 +0100: > On 2 Apr 2008, at 02:03, Steve Chervitz wrote: > >> You're right. The bifurcation that has me concerned is really: >> >> DAS/1 --> genome & protein >> DAS/2 --> genome only >> >> Though DAS/2 can be used for protein sequence annotations, it has >> so far >> mainly (only?) been used for serving genome annotations. The >> retrieval DAS/2 >> spec itself is very genome-centric and could be updated/generalized to >> include protein examples. >> >> So I guess it's not so much a bifurcation but a disparity in level of >> support and usage patterns that distinguishes the 1 vs 2 versions >> of the >> spec, and this disparity seems destined to ensure the co-existence >> incompatible DAS 1 & 2 clients and servers. >> >> DAS/1 has a large installed base of servers and clients and the DAS/ >> 1 spec >> has more extensive support for protein sequence annotations. >> >> DAS/2 has some enhancements including support for alternative >> formats and an >> incipient writeback (annotation editing) facility. Usage as so far >> been >> confined to genome browsers. >> > > One important thing in my mind is that "writeback" is a really huge > additional concept that carries so many semantic issues which many servers > just cannot implement (for sure, Ensembl is not going to be implementing > a writeback server in the near term). > > For me, DAS's sweet spot is a "dumb server, smart client" mode, where > there a few clients and many servers, and servers are easy to put up > etc. Write back requires smart(er) servers. > > So I would also advocate splitting the DAS/2 discussing into "read-only" > and "read-write" scenarios. I know this is implicitly there but I would > argue worth formalising at a higher level, and would stress the servers > can still be dumb.
Agreed. Lots of issues involved with writeback. That portion of the DAS/2 spec is clearly segregated from the retrieval portion, and was still in the experimental stage when it was last worked on (~summer 2007). The future of this aspect of the spec isn't so clear to me. It would take a significant amount of developer time and community interest to make it fly. So I would de-emphasize the writeback stuff when discussing what DAS/2 brings to the table. Seems to me to be more important to first address how to best align the DAS/2 - DAS/1 work within the read-only world before getting into writeback, and only then if there's significant interest & funding. > It is also worth noting that DAS/1 has evolved as a "de facto", working > standard around the hinxton campus and we've really made good use and > built up momentum around it. It takes a long time bizarrely for people > to pick up technology and use it, and I want to make sure DAS/2 > discussion doesn't scare the growing set of people picking up and using > DAS/1 - which is very useful, pragmatic and critically has working > servers and clients now. The growth in DAS users is a good thing and should be encouraged since it helps realize the distributed annotation vision, and hooks up more people with more data. Whether DAS/1 or DAS/2 or DAS/n is used under the hood is a technical issue. One can always construct smarter clients that know how to speak whatever DAS flavor, though that's not necessarily a good long-term solution. It's clear that there is evolutionary pressure on the DAS spec, both in the existence of DAS/2 and in the modifications to DAS/1. What's less clear is how to channel these forces so that the innovative ideas on both sides of the pond can come together. Steve >> [ BTW: there isn't a good document describing these DAS/2 >> enhancements and >> how DAS/2 diverges from DAS/1. I'll try and put something up on the >> biodas.org wiki about this. In the meantime, there is a brief >> summary here: >> http://biodas.org/documents/das2/das2_protocol.html ] >> >> I agree with your ultimate goal for DAS. Getting there will take some >> effort, but promoting that vision is a important step. Resuming the >> conference calls as Ann mentioned will help. >> >> Cheers, >> Steve >> >> >>> From: Andreas Prlic <[EMAIL PROTECTED]> >>> Date: Tue, 1 Apr 2008 22:37:18 +0100 >>> To: Steve Chervitz <[EMAIL PROTECTED]> >>> Cc: DAS/2 Discussion <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> >>> Subject: Re: [DAS] DAS workshop presentations >>> >>> Hi Steve, >>> >>> I don't really see the bifurcation between protein and genome DAS >>> that you mention. We use DAS/1 nicely for both genome and protein >>> clients. >>> Also the DAS/1 extensions have been designed with being able to work >>> with both worlds. >>> >>> The real bifurcation I see is the geographical split between the US >>> and Europe as DAS is used in different ways here and there. Europe is >>> very much organized around DAS/1. >>> >>> The ultimate goal of DAS is to separate data provision from >>> visualization. As such a goal for us all should be that a DAS source >>> can be shown in a similar fashion in all (genome) browsers. At the >>> present this is not possible. In order to achieve such a standard it >>> still will require a lot of collaboration, standardization and >>> convincing between the different client and server-providers. >>> >>> Andreas >>> >>>> >>>> There are perhaps are a handful of sites in the U.S. that have (or >>>> are planning to) deploy DAS/2 servers in production environments, >>>> though I don't have any official numbers. We've been tracking them >>>> here: http://www.biodas.org/wiki/DAS/2#DAS.2F2_Servers >>>> >>>> Since the DAS/2 retrieval spec stabilized, there has been some work >>>> on the writeback portion of the spec (for creating/editing >>>> annotations), but that has not been finalized. As far as I know, it >>>> hasn't progressed since the original grant for DAS/2 spec >>>> development ended in Oct 2007. There was some good work done on the >>>> writeback spec as well as some early implementations and it would >>>> be good to see it congeal eventually. >>>> >>>> I know there's been continued development of the DAS/1 spec: http:// >>>> www.dasregistry.org/spec_1.53E.jsp >>>> >>>> This has stirred up some thoughts I've been chewing on recently >>>> regarding where DAS is headed. Comments welcome. Here goes: >>>> >>>> While it would be nice to see a single DAS spec to simplify the >>>> lives of people writing clients and to unite development efforts, I >>>> wonder if this is reasonable expect yet. The genome vs. protein >>>> annotation bifurcation that we have now may reflect a natural >>>> division that might be difficult to unite into a single spec. >>>> However, I think the current state is more a matter of the large >>>> amount of inertia behind DAS/1 -- the shear number of existing DAS/ >>>> 1 servers and DAS/1 client code out there -- coupled with the fact >>>> that DAS/2 is not backward compatible. Code doesn't rewrite itself. >>>> >>>> I know one of the early roadblocks for getting protein annotation >>>> support into DAS/2 was that protein features were not described by >>>> the Sequence Ontology (SO), but this has since been rectified. I've >>>> learned that since August 2007, all of the protein features in the >>>> bioSapiens project have been incorporated into SO ( http:// >>>> www.dasregistry.org/extension_ontology.jsp ). >>>> >>>> The DAS/1 -> DAS/2 proxy adapter that Andrew Dalke worked on >>>> ( http://lists.open-bio.org/pipermail/das2/2006-October/ >>>> 000268.html ) could help with DAS/2 migration. But that project >>>> never fully matured. Continued development of DAS/1 could >>>> complicate the development of such an adapter. >>>> >>>> Regarding the genome-vs-protein DAS split, certainly there are many >>>> commonalities at the level of annotations along a linear sequence >>>> which apply equally well to nucleotide and protein sequences. Maybe >>>> it would help to have a partitionable DAS spec to separate core >>>> things from extensions that are nucleotide or protein-specific? >>>> This could help isolate the more stable, low-level things from >>>> higher-level functionality that is specific to nucleotide or >>>> protein sequences, and could evolve without impacting the core >>>> spec. A given DAS server would be compliant with the core and >>>> either the genome or protein specific extensions of the spec. >>>> >>>> Cheers, >>>> Steve >>>> >>>> >>>>> From: Ann Loraine <[EMAIL PROTECTED]> >>>>> Date: Thu, 27 Mar 2008 10:04:21 -0400 >>>>> To: Andreas Prlic <[EMAIL PROTECTED]> >>>>> Cc: Steve Chervitz <[EMAIL PROTECTED]>, DAS/2 >>>>> Discussion >>>>> <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> >>>>> Subject: Re: [DAS] DAS workshop presentations >>>>> >>>>> Regarding the funding issue: >>>>> >>>>> NSF is awarding me some funds that will support developing a >>>>> Distributed Annotation Server to use for Arabidopsis data. (It is >>>>> through the 2010 program.) >>>>> >>>>> DAS is not the main focus of the grant, but will play a big part >>>> in what we >>>>> do. >>>>> >>>>> I am also working on arranging a number we could use to carry on >>>>> the >>>>> DAS conference calls. >>>>> >>>>> Previously they were hosted by Affymetrix, but I think we can >>>> probably >>>>> do this for the group here at Charlotte. (I'm at University of >>>>> North >>>>> Carolina Charlotte.) >>>>> >>>>> All the best, >>>>> >>>>> Ann >>>>> >>>>> On Thu, Mar 27, 2008 at 9:35 AM, Andreas Prlic <[EMAIL PROTECTED]> >>>> wrote: >>>>>> Hi Steve, >>>>>> >>>>>> We did not have an "official" slot for DAS/2, partially >>>>>> because we >>>>>> did not have any DAS/2 related speakers. In the informal >>>> discussions >>>>>> some concern re DAS/2 was raised, >>>>>> particularly about backwards compatibility issues and regarding >>>>>> whether DAS/2 is still active and funded. >>>>>> >>>>>> Cheers, >>>>>> Andreas >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On 27 Mar 2008, at 01:24, Steve Chervitz wrote: >>>>>> >>>>>>> Thanks Andreas. Looks like you had a good workshop. >>>>>>> Was there any discussion pertaining to DAS/2? >>>>>>> >>>>>>> Steve >>>>>>> >>>>>>>> From: Andreas Prlic <[EMAIL PROTECTED]> >>>>>>>> Date: Wed, 26 Mar 2008 16:01:26 +0000 >>>>>>>> To: <[EMAIL PROTECTED]> >>>>>>>> Subject: [DAS] DAS workshop presentations >>>>>>>> >>>>>>>> Hi, >>>>>>>> >>>>>>>> We recently were hosting a DAS workshop at the Genome Campus in >>>>>>>> Hinxton, U.K.. >>>>>>>> >>>>>>>> In case you are interested, the presentations of the first day >>>> (DAS - >>>>>>>> client developers day) are available from: >>>>>>>> http://www..dasregistry.org/course.jsp >>>>>>>> >>>>>>>> The slides used for the second day (hands-on) are available >>>>>>>> from: >>>>>>>> http://www.biodas.org/wiki/DASworkshop200802 >>>>>>>> >>>>>>>> Cheers, >>>>>>>> Andreas >>>>>>>> >>>>>>>> >>>> -------------------------------------------------------------------- >>>> - >>>>>>>> -- >>>>>>>> >>>>>>>> Andreas Prlic Wellcome Trust Sanger Institute >>>>>>>> Hinxton, Cambridge CB10 1SA, UK >>>>>>>> +44 (0) 1223 49 6891 >>>>>>>> >>>>>>>> >>>> -------------------------------------------------------------------- >>>> - >>>>>>>> -- >>>>>>>> >>>>>>>> -- >>>>>>>> The Wellcome Trust Sanger Institute is operated by Genome >>>> Research >>>>>>>> Limited, a charity registered in England with number 1021457 >>>> and a >>>>>>>> company registered in England with number 2742969, whose >>>> registered >>>>>>>> office is 215 Euston Road, London, NW1 2BE. >>>>>>>> _______________________________________________ >>>>>>>> DAS mailing list >>>>>>>> [email protected] >>>>>>>> http://lists.open-bio.org/mailman/listinfo/das >>>>>>> >>>>>> >>>>>> >>>> -------------------------------------------------------------------- >>>> -- >>>> - >>>>>> >>>>>> Andreas Prlic Wellcome Trust Sanger Institute >>>>>> Hinxton, Cambridge CB10 1SA, UK >>>>>> +44 (0) 1223 49 6891 >>>>>> >>>>>> >>>> -------------------------------------------------------------------- >>>> -- >>>> - >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> The Wellcome Trust Sanger Institute is operated by Genome >>>> Research >>>>>> Limited, a charity registered in England with number 1021457 >>>> and a >>>>>> company registered in England with number 2742969, whose >>>> registered >>>>>> office is 215 Euston Road, London, NW1 2BE. >>>>>> _______________________________________________ >>>>>> DAS mailing list >>>>>> [email protected] >>>>>> http://lists.open-bio.org/mailman/listinfo/das >>>>>> >>>> This transmission is intended for the sole use of the individual >>>> and entity to whom it is addressed, and may contain information >>>> that is privileged, confidential and exempt from disclosure under >>>> applicable law. You are hereby notified that any use, >>>> dissemination, distribution or duplication of this transmission by >>>> someone other than the intended addressee or its designated agent >>>> is strictly prohibited. If you have received this transmission in >>>> error, please notify the sender immediately by reply to this >>>> transmission and delete it from your computer. Thank You. >>>> Affymetrix, Inc. >>> >>> --------------------------------------------------------------------- >>> -- >>> >>> Andreas Prlic Wellcome Trust Sanger Institute >>> Hinxton, Cambridge CB10 1SA, UK >>> +44 (0) 1223 49 6891 >>> >>> --------------------------------------------------------------------- >>> -- >>> >>> >>> >>> >>> -- >>> The Wellcome Trust Sanger Institute is operated by Genome Research >>> Limited, a charity registered in England with number 1021457 and a >>> company registered in England with number 2742969, whose registered >>> office is 215 Euston Road, London, NW1 2BE. >> >> >> ------------------------------------------------------------ >> >> This transmission is intended for the sole use of the individual >> and entity to whom it is addressed, and may contain information >> that is privileged, confidential and exempt from disclosure under >> applicable law. You are hereby notified that any use, >> dissemination, distribution or duplication of this transmission by >> someone other than the intended addressee or its designated agent >> is strictly prohibited. If you have received this transmission in >> error, please notify the sender immediately by reply to this >> transmission and delete it from your computer. Thank You. >> Affymetrix, Inc. >> _______________________________________________ >> DAS mailing list >> [email protected] >> http://lists.open-bio.org/mailman/listinfo/das > _______________________________________________ DAS mailing list [email protected] http://lists.open-bio.org/mailman/listinfo/das
