[DAS] DAS projects/Future of DAS

Jonathan Warren jw12 at sanger.ac.uk
Thu Nov 24 17:19:37 UTC 2011


Hi Manuel

Thanks again for the input ;)

I agree with Andy that a generic DAS upload/server is going to be  
inherently complicated and will be limited by the upload across the  
internet. To a large degree the DAS system came about to stop large  
amounts of data needing to be uploaded or downloaded over the net.
However I can see that having DAS "nodes" such as the EBI and Sanger  
etc with tailor made upload user interfaces and servers for specific  
data types is a reasonable solution/addition to the DAS system. To  
this end, I'm already more than half way to a solution for your needs  
(which doesn't need a database) which we could use, subject to Sanger  
approval. This solution could also be made more generic for other data  
sources as well with a specific interface developed for each.

As to the future of DAS - I believe that over the last 3 years we have  
improved the DAS protocol and many of the associated implementations  
so that we now have dispensed with many if not all of the previous  
criticisms people have had of the DAS system (1.6E spec):
Validation has improved the level of conformity to the DAS spec and  
new servers (MyDAS and Proserver) now behave in the same way for both  
requests and responses.
You can now search DAS sources (MyDAS).
Next feature capability (in the spec and Proserver).
You can have alternative content that will require less bandwidth e.g  
JSON (the Registry serves this already and soon MyDAS  servers and  
proserver hopefully).
We have writeback servers (MyDAS)-(already implemented at the EBI for  
proteins and an example server available soon that will accept posts  
and puts at the Sanger for genomic sources).

Really we NEED the community to come together and put data up using  
these new servers and for major clients such as Ensembl to start  
supporting 1.6 spec servers and the newer "Extended" features. The  
Dalliance browser by Thomas Down proves how fast a DAS client can be.  
I believe there is a lot more potential for the DAS system and it's  
still a good solution for today's data distribution needs.

Cheers

Jonathan.



On 19 Nov 2011, at 15:58, Manuel Corpas wrote:

>> Having said all this, I am a little confused about what you are  
>> trying to achieve. In your first mail you said you wanted to create  
>> sources via an API, in the second you say you want to do it via a  
>> click. Obviously the requirements for both are very different.
>
>
> Both API and 1-click DAS source creation would be extremely helpful in
> my view. The fact that this functionality is not available is
> seriously affecting many of our users' ability to create DAS sources
> with their 23andMe genotypes.
>
> The fact that these facilities do not exist has stopped new potential
> users from utilizing DAS. If DAS is truly going to survive as a
> standard, automatic creation of data sources needs to be easier.
>
> Manuel
>
>
> Manuel Corpas, PhD
> Tel:      +44.122349.2372
> Web:    http://manuelcorpas.com/about/
> Twitter: @manuelcorpas
>
>
>
> On 18 November 2011 23:33, Andy Jenkinson <andy.jenkinson at ebi.ac.uk>  
> wrote:
>> Hi Manuel,
>>
>> It would be nice to be able to create a DAS source from any type of  
>> data you happen to have with a click or two, but I don't think it  
>> is realistic. Even in this email you have just told me what all the  
>> columns mean, what the assembly is, what kind of file it is. Any  
>> application would need to know the same things (and more).
>>
>> That is not to say that it is difficult to build something to let  
>> you do this if it is specifically designed for the exact type of  
>> data you are using, just that it does not already exist and so you  
>> have to actually create it. Either MyDas or ProServer would seem to  
>> offer you a starting point to do that, but only a starting point.  
>> EasyDAS is the closest thing to what you want but obviously it has  
>> to cater for any type of data so has to ask you a lot more  
>> questions. Its web-based architecture obviously limits the size of  
>> data files you can process quickly too, but that is the trade off  
>> you make by not needing an Internet-visible web server of your own  
>> to run a DAS server from. I daresay if you wanted to create  
>> something that an individual can use to make a DAS source from  
>> their personal BED/VCF file then it would have to be web based,  
>> will always be restricted by the speed of the Internet, but the  
>> interface could be much simpler than EasyDAS and a database might  
>> not be needed (EasyDAS loads file contents into a database to  
>> standardise them, which slows things down).
>>
>> Having said all this, I am a little confused about what you are  
>> trying to achieve. In your first mail you said you wanted to create  
>> sources via an API, in the second you say you want to do it via a  
>> click. Obviously the requirements for both are very different.
>>
>> Cheers,
>> Andy
>>
>> On 18 Nov 2011, at 15:24, Manuel Corpas wrote:
>>
>>> Hi Andy,
>>>
>>> thanks for the info. Having a bed DAS adaptor is part of the  
>>> problem,
>>> the other is not having to worry about having to deal with the DAS
>>> server directly. easyDAS manages to do this but unfortunately it is
>>> not obvious for people who do not know DAS how to operate it. Also  
>>> if
>>> the file is very big and the connection slow it can take up to an  
>>> hour
>>> to create a DAS source.
>>>
>>> Wouldn't it be nice to create a DAS source just with one click or  
>>> two?
>>>
>>> Please see below a snippet of a few SNPs in my chromosome 16 just as
>>> you would get them from 23andMe (NCBI36 assembly; columns mean
>>> SNP_id/chr/position/genotype).
>>>
>>> Cheers,
>>> Manuel
>>>
>>> rs7763        16      544555  TT
>>> rs763158      16      546105  GG
>>> rs7190878     16      549131  AG
>>> rs4984890     16      552699  CT
>>> rs710925      16      573355  AG
>>> rs2017567     16      577213  CT
>>> rs4144003     16      585969  CT
>>> rs7190358     16      590789  AG
>>> rs7203694     16      592942  AG
>>> rs11248940    16      595687  TT
>>> rs7204088     16      601143  TT
>>> rs4984677     16      611683  AG
>>> rs9929621     16      619413  CT
>>> rs11642546    16      641657  CC
>>> rs3752496     16      650256  TT
>>> rs2301426     16      651906  GG
>>> rs1044662     16      655061  CC
>>> rs9934288     16      656288  AC
>>> rs3752493     16      657524  TT
>>> rs1139897     16      660987  GG
>>> rs1045763     16      664085  CC
>>> rs3830140     16      665336  AA
>>> rs8056588     16      666190  CC
>>> rs6597        16      671726  TT
>>>
>>>
>>> Manuel Corpas, PhD
>>> Tel:      +44.122349.2372
>>> Web:    http://manuelcorpas.com/about/
>>> Twitter: @manuelcorpas
>>>
>>>
>>>
>>> On 18 November 2011 15:14, Andy Jenkinson  
>>> <andy.jenkinson at ebi.ac.uk> wrote:
>>>> Hi Manuel,
>>>>
>>>> Since 2008 ProServer has had a BED format SourceAdaptor (called  
>>>> bed12, as it is intended to work with the 12-field BED format).  
>>>> It also supports Hydras, which are modules that are designed to  
>>>> automatically create DAS sources from a single config without  
>>>> restarting the server. This is how EasyDAS works with ProServer:  
>>>> there is one SourceAdaptor, and a Hydra to scan a relational  
>>>> database for new data.
>>>>
>>>> I don't know what 23andme's data looks like, but the addition of  
>>>> a Hydra to scan directories for new files and automatically make  
>>>> them available as DAS sources would seem to be a trivial piece of  
>>>> work. I daresay a VCF adaptor would also be fairly easy,  
>>>> especially if there is a Perl API of some sort (BioPerl?).
>>>>
>>>> Cheers,
>>>> Andy
>>>>
>>>> On 17 Nov 2011, at 17:11, Manuel Corpas wrote:
>>>>
>>>>> Dear Jonathan,
>>>>>
>>>>> I hope you do not mind me copying the DAS list in this email, as  
>>>>> we
>>>>> would be very keen to gather interest in the community regarding  
>>>>> DAS
>>>>> applications to whole genomes.
>>>>>
>>>>> We are interested in exploring DAS in the context of genomic  
>>>>> variants
>>>>> (SNPs, indels, CNVs) from personal genomes plus their  
>>>>> integration with
>>>>> relevant sources (genes, variation data, phenotypes).
>>>>>
>>>>> Currently we have done a lot of work with 23andMe (whole-genome)
>>>>> genotypes but now we are expecting to extend our efforts further  
>>>>> to
>>>>> exome data. A critical tool we are currently missing is one that
>>>>> allows automatic creation of DAS sources via an API directly  
>>>>> from bed
>>>>> format (used by 23andMe) or vcf (1000genomes).
>>>>>
>>>>> Anyone interested in discussing these topics please let me know.
>>>>>
>>>>> Kind regards,
>>>>> Manuel
>>>>>
>>>>> Manuel Corpas, PhD
>>>>> Tel:      +44.122349.2372
>>>>> Web:    http://manuelcorpas.com/about/
>>>>> Twitter: @manuelcorpas
>>>>>
>>>>>
>>>>>
>>>>> On 17 November 2011 12:11, Jonathan Warren <jw12 at sanger.ac.uk>  
>>>>> wrote:
>>>>>> Hi
>>>>>>
>>>>>> As the 2012 DAS workshop is coming up at the end of February we  
>>>>>> would like
>>>>>> to hear from people using DAS.
>>>>>> We would be really grateful to receive just a short email from  
>>>>>> anyone using
>>>>>> DAS or developing DAS with a brief summary about their project  
>>>>>> and how DAS
>>>>>> fits in, especially if you have not spoken at the DAS workshops  
>>>>>> at any time.
>>>>>>
>>>>>> Please also say if you would be interested in giving a short  
>>>>>> presentation at
>>>>>> the workshop in February even if you are not sure if you could  
>>>>>> make it.
>>>>>> Previous years the presentations have been 15 minutes with 5  
>>>>>> minutes for
>>>>>> questions - however this year we intend to be more flexible and  
>>>>>> so if you
>>>>>> would prefer to give a "lightning talk" of just 5 minutes to  
>>>>>> update people
>>>>>> or give them a brief overview that will be fine. Links to the  
>>>>>> previous years
>>>>>> talks can be found here http://www.biodas.org/wiki/DASWorkshop2011#Day_2
>>>>>>
>>>>>> I must emphasise - please give us a summary even if you are not  
>>>>>> interested
>>>>>> in giving a talk as we would like to know what is going on out  
>>>>>> there and we
>>>>>> promise not to hound you to give a talk :)
>>>>>>
>>>>>> Thanks in advance
>>>>>>
>>>>>> The Sanger/EBI DAS people.
>>>>>>
>>>>>>
>>>>>> Jonathan Warren
>>>>>> Senior Developer and DAS coordinator
>>>>>> blog: http://biodasman.wordpress.com/
>>>>>> jw12 at sanger.ac.uk
>>>>>> Ext: 2314
>>>>>> Telephone: 01223 492314
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> The Wellcome Trust Sanger Institute is operated by Genome  
>>>>>> ResearchLimited, a
>>>>>> charity registered in England with number 1021457 and acompany  
>>>>>> registered in
>>>>>> England with number 2742969, whose registeredoffice is 215  
>>>>>> Euston Road,
>>>>>> London, NW1 2BE._______________________________________________
>>>>>> DAS mailing list
>>>>>> DAS at lists.open-bio.org
>>>>>> http://lists.open-bio.org/mailman/listinfo/das
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> DAS mailing list
>>>>> DAS at lists.open-bio.org
>>>>> http://lists.open-bio.org/mailman/listinfo/das
>>>>
>>>>
>>
>>

Jonathan Warren
Senior Developer and DAS coordinator
blog: http://biodasman.wordpress.com/
jw12 at sanger.ac.uk
Ext: 2314
Telephone: 01223 492314









-- 
 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. 



More information about the DAS mailing list