Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

searching for unpublished GO-Annotated sequences for application-valuation

David Martin david at compbio.dundee.ac.uk
Tue Mar 15 07:06:19 PST 2005


On 15/3/05 2:08 pm, "zehetner at molgen.mpg.de" <zehetner at molgen.mpg.de> wrote:

> Hallo Stefan,
> 
> I have developed an annotation prediction system which is based on a modified
> OntoBlast tool (OntoBlast function: from sequence similarities directly to
> potential functional annotations by ontology terms, Nucleic Acids Research,
> 2003, Vol. 31, No. 13, pp 3799-3803) and a prototype implementation for
> malaria
> species (Anopheles and Plasmodium) can be found at http://malariabase.org/
> This predicts many not yet available functional annotations (using GO terms)
> for
> genes from these species. Although they are not manually but computationally
> derived, those with low weighting numbers (< 1.0e-100) are in general very
> reliable. You are welcome to use them to evaluate your system if you wish.
> If you have any questions please let me know.

We also used the malaria data set as part of the evaluation of GOtcha
(http://wwww.compbio.dundee.ac.uk/). In our experience you are quite right
to exclude the IEA terms from the analyses as these will 'muddy the waters',
giving a measure of how well one retrieves functions by similarity
searching. For a starting point I would suggest taking the GO-seq data and
excluding all the IEA graded terms. This should provide a suitably large
data set.

Of particular interest to me would be exactly how you propose to measure the
accuracy of the tool you are developing. I am not yet convinced that there
is any *obviously best* way to perform such an analysis.

.d
-- 
David Martin PhD
Post-Genomics and Molecular Interactions Centre
University of Dundee
http://www.compbio.dundee.ac.uk/


> 
> Best wishes,
> Günther Zehetner
> 
> MPI for Molecular Genetics, Berlin
> 
> 
> Quoting Stefan Goetz <goetz_stefan at gmx.de>:
> 
>> Hello everybody,
>> 
>> we are developing a new GO-Tool which includes gene function
>> prediction/annotation based on sequence similarities. The evaluation of
>> prediction power involves the comparison of our annotation resutls with
>> manually annotated data.
>> 
>> For this purpose we would like to use a set of  manually GO annotated
>> sequences (nuclein/protein) not yet included in public databases (e.g.
>> NR from NCBI). By this we try to minmize the problem that the sequences
>> of the test-set may have influenced the IEA annotation of the sequences
>> in the search set. We want to stress that our interest is merely the
>> validation of our tool and not the specific biological meaning of a
>> particular data set, and that we would treat confidentially any
>> unpublished data.
>> 
>> We will be very pleased if someone could supply as with such kind of
>> sequences.
>> 
>> Regards, Stefan Götz
>> 
>> IVIA - BET
>> Universidad Politecnica de Valencia
>> http://gim.upv.es/
>> http://www.ivia.es/
>> 
>> 
>> 
>> 
>> --
>> This message is from the GOFriends moderated mailing list.  A list of public
>> announcements and discussion of the Gene Ontology (GO) project.
>> Problems with the list?           E-mail: owner-gofriends at geneontology.org
>> Subscribing   send   "subscribe"   to   gofriends-request at geneontology.org
>> Unsubscribing send   "unsubscribe"  to  gofriends-request at geneontology.org
>> Web:          http://www.geneontology.org/
>> 
> 
> 
> 
> 
> 
> 
> --
> This message is from the GOFriends moderated mailing list.  A list of public
> announcements and discussion of the Gene Ontology (GO) project.
> Problems with the list?           E-mail: owner-gofriends at geneontology.org
> Subscribing   send   "subscribe"   to   gofriends-request at geneontology.org
> Unsubscribing send   "unsubscribe"  to  gofriends-request at geneontology.org
> Web:          http://www.geneontology.org/



--
This message is from the GOFriends moderated mailing list.  A list of public
announcements and discussion of the Gene Ontology (GO) project.
Problems with the list?           E-mail: owner-gofriends at geneontology.org
Subscribing   send   "subscribe"   to   gofriends-request at geneontology.org
Unsubscribing send   "unsubscribe"  to  gofriends-request at geneontology.org
Web:          http://www.geneontology.org/



More information about the go-friends mailing list