Hi Arek,
Thank you for your offer of help.
Here are some examples, and if you can do something with them that would
be great:-
EMAGE text expression annotation dataset has attribute "Atlas ontology
ID of tissue (text annotation) " ie the ID of the anatomy ontology
component that has been annotated, but users would like to know the
component name, component path, developmental stage
and so on for that ID. These attributes are in the dataset Anatomy
components (component details), where the common attribute is ID (timed
stages).
From the datasets Anatomy components (component details), and Anatomy
components (including subcomponents), users would like to know if the
anatomy components they are interested in have gene expression annotated
in EMAGE. Details of the annotation (gene, strength, pattern etc) are in
the EMAGE text expression annotation dataset. As above, the common
attribute is the ontology ID.
The EMAGE submission dataset contains data regarding an individual assay
submitted to EMAGE. However it does not have details of the detection
reagents (probes) used, these details are in the EMAGE experiment
dataset. Users would be interested in the probe details for EMAGE
submissions that they have found. There are several common attributes
between these datasets but the most obvious one would be the EMAGE ID.
In the other direction, users might search for probes in the EMAGE
experiment dataset and want to know more details about the EMAGE
submission.
The EMAGE submission dataset does not have details of spatial expression
annotation, this is in the EMAGE spatial expression dataset. It is
likely that users would want to know if a submission they had found has
spatial expression annotated.
In the other direction, users looking for spatial expression might want
to know more details about the submission. The EMAGE ID would be the
common attribute again here.
I hope that this makes sense but obviously let me know if you need more
information.
Regards,
Peter.
On 27/09/2011 15:38, Arek Kasprzyk wrote:
Hi Peter,
i did not make myself clear. I was not talking about bringing user
defined linking back in. I was more thinking along the lines that
there are typically certain datasets and that your users should like
to link in order to do their favourite multi-dataset queries. If you
tell us what they are we'll be able to preconfigure the interface such
that these queries become possible again without users being forced to
do the GUI gymnastics as they used to do in 0.7.
We just need concrete examples of multi-datasets queries that your
users typically do and we could preconfigure them for you if it makes
sense.
As i mentioned in my previoius email, in the future you will be able
to do such things on the central portal yourself but this is ... in
the future.
For now we need an email from you with concrete examples so we can see
if we can help you and keep your users happy :)
a
On Tue, Sep 27, 2011 at 9:01 AM, Peter Stevenson
<[email protected] <mailto:[email protected]>>
wrote:
Hi Arek,
I just meant linking in the v0.7 sense of using
importables/exportables in order for the user to get attributes
from separate datasets. I can give details of the
importables/exportables I used if that's useful. However it sounds
from the replies from Jonathan and Syed that its possibly not that
straightforward to bring that type of configuration into a v0.8 model?
Regards,
Peter.
On 09/26/11 19:14, Arek Kasprzyk wrote:
2. User-definable linking is not currently implemented
in 0.8.
This is a pity particularly for us because the EMAGE
implementation relies
on this feature quite heavily.
Hi Peter,
it would be helpful if could suggest what type of linking you
are after.
This can be pre-configured for you on BMC so your users would
not be
affected. In the future we'll be able to do add your source
yourself and
configure it in the Central Registry through our
configuration tool
MConfigurator but for now you need to do via email :)
a
On 09/26/11 15:12, Jonathan Guberman wrote:
Hello Peter,
1. EMAP anatomy ontology database is included: it is
under the "Search by
organism" tab, under vertebrates.
2. User-definable linking is not currently implemented
in 0.8.
3. There is a problem with the 0.7 configuration that
prevents these
filters and attributes from being imported into 0.8:
in your configuration
XML (available at http://biomart.emouseatlas.**
org/biomart/martservice?type=**configuration&dataset=**EXPRESSION_TEXT_TB<http://biomart.emouseatlas.org/biomart/martservice?type=configuration&dataset=EXPRESSION_TEXT_TB
<http://biomart.emouseatlas.org/biomart/martservice?type=configuration&dataset=EXPRESSION_TEXT_TB>>)
there is one MainTable/key pair listed at the top of
the XML:
<MainTable>EXPRESSION_TEXT_TB_**_EXPRESSION_TEXT_TB__main</**MainTable>
<Key>OID_109_key</Key>
However, certain attributes and filters refer to other
keys. E.g.:
<FilterDescription displayName="Experiment ID"
displayType="text"
field="oid_108"
hideDisplay="true"**internalName="oid_108_filter"
key="OID_108_key" legal_qualifiers="like" qualifier="like"
tableConstraint="main"/>
And
<AttributeDescription default="true"
displayName="Emage ID"
field="accession_1037"
internalName="accession_1037"**key="OID_1037_key"
linkoutURL="http://www.**emouseatlas.org/emagewebapp/**
<http://emouseatlas.org/emagewebapp/**>
pages/emage_entry_page.jsf?id=**EMAGE:%s"
maxLength="20"tableConstraint=*
*"main"/>
Since these filters and attributes refer to keys about
which no
information is provided in the XML, they can not be
recreated by 0.8 and are
ignored.
4. and 5. BioMart 0.8 does not currently have these
capabilities, but they
are planned for future releases.
Sincerely,
Jonathan Guberman, PhD
Application Programmer
Ontario Institute for Cancer Research
MaRS Centre, South Tower
101 College Street, Suite 800
Toronto, Ontario, Canada M5G 0A3
Tel: 647-260-7818 <tel:647-260-7818>
Toll-free: 1-866-678-6427 <tel:1-866-678-6427>
www.oicr.on.ca
<http://www.oicr.on.ca><http://www.**oicr.on.ca
<http://oicr.on.ca><http://www.oicr.on.ca>>
On 2011-09-26, at 9:27 AM, Peter Stevenson wrote:
I was looking at http://central.biomart.org/ after
reading the recent
paper in Database.
It's nice work, looks good, and I'm obviously happy to
see the EMAGE
biomart included in there.
However I have some quibbles about how the EMAGE
interface has been
represented on there...
1. The EMAP anatomy ontology database and its 2
datasets are not included.
2. There is no mechanism to do linked dataset queries.
3. EMAGE text expression annotation dataset has
several groups of filters
and attributes missing.
4. There are no default attributes pre-selected.
5. There is no select all/deselect all in groups of
attributes.
I don't know if some of these are technical
limitations with representing
a v0.7 implementation in
v0.8? I recognise some issues from when I tried to
port to a previous v0.8
release candidate. Are
there changes at this end that I could make to help?
Regards,
Peter.
--
--
Peter Stevenson | [email protected]
<mailto:[email protected]>
EMAP Computer Support | http://www.emouseatlas.org/emage
MRC Human Genetics Unit | Phone: +44 (0)131 332 2471
<tel:%2B44%20%280%29131%20332%202471>
Edinburgh EH4 2XU. U.K. |
_______________________________________________
Users mailing list
[email protected]
https://lists.biomart.org/mailman/listinfo/users
