[DAS] DAS writeback

Gustavo Salazar gsalazar at cs.uct.ac.za
Thu Apr 30 10:22:36 UTC 2009


Hello all,

After our discussion during the DAS workshop, I've been working 
implementing and redefining some things for the writeback, and that's 
why I haven't wrote the Writeback protocol proposal as I promised, so 
I'm sorry about it but I hope that when I put it, it will be of your 
interest.
The main change I'm working is that now the input format is the same 
output (i.e. DASGFF), and the action to do is the HTTP method that is 
been used (POST:Create, PUT: Update, DELETE:delete, GET:query).
The other feature that I'm working on, is that the response of the 
writeback server will be a DASGFF file that contains the information 
that was succesfully processed.
I will try to put a document that describe this in more details in the 
next weeks. I want to first finish the implementation to publish both 
things together.

Now, I have a question that I need to clarify to finish both document 
and implementation. I hope you can help me guys...
* Is there any way to know really identify the segment from the GFF 
file? For instance the response  of this server
http://das.ensembl.org/das/ens_35_segdup_washu/features?segment=17:13650001,14649999

is something like:

<?xml version="1.0" standalone="yes"?>
<?xml-stylesheet type="text/xsl" href="features.xsl"?>
<!DOCTYPE DASGFF SYSTEM "http://www.biodas.org/dtd/dasgff.dtd">
<DASGFF>
  <GFF version="1.01" 
href="http://das.sanger.ac.uk:80/das/ens_35_segdup_washu/features">
    <SEGMENT id="17" version="1.0" start="13650001" stop="14649999">
        <FEATURE id="17:13864005-13874846" label="17:13864005-13874846">

          ....


And the id of the segment is 17 because is from the chromosome 17, but 
it could be from any organism. I though in defining the ID as an URL 
formed with the http value of the GFF element plus the segment id, 
however it looks to me that it will be attached to the feature source 
and no to the reference; and therefore 2 features coming from different 
annotation servers but for the same segment will be hard to group in the 
same segment.
This is very important, because the writeback server should be able to 
behave as an annotation server and if a client request features of a 
segment the wb server should return all the features there without 
discriminate by original source. This is really simple for proteins (My 
really interest) because the segment has the uniprot id, so i am sure 
what is the resource that the client want, but for chromosome info 
doesn't look that easy
I hope you understand my problem, and somebody can help me.

Regards,

Gustavo Salazar



More information about the DAS mailing list