Dear Drosophila database masters:
I have a few general questions regarding what is being generated through the
database and if there is a 'better' means of assessing some of the results.
I am a protein chemist/structural biologist working on three drosophila
proteins. The first protein is Roundabout-1 (Robo-1). This protein has been
well characterized and demonstrates the common motifs found in cell adhesion
proteins, and has thus been characterized as a cell adhesion protein in your
database. This characterization is apparently wrong however because the Ig
domains in robo are known to bind the slit protein and have no impact on cell
adhesion. This leads to my first question about how experimental data is
being incorporated into the databases that we currently have access to. I do
understand that the task is difficult given the large number of sequences to
be used. I feel some of this could be solved by integration with databases
such as provided at the interactive fly: This would provide some experimental
links and possibly help to identify homologous proteins.
http://sdb.bio.purdue.edu/fly/aimain/1aahome.htm Is there some work in
progress to link experimental data to the sequences? How or when can this be
accessed?
My second question is a bit more difficult to answer, but please provide any
assistance that you think will be useful. As for the robo proteins, there are
three members of the family robo-1,-2, and-3. Robo-1 has been characterized,
but the other two have not. When the database is searched, robo-2 and robo-3
are found, but the identified cDNAs only include the extracellular portions of
the proteins-which are composed of homologous Ig domains. The cytoplasmic
tails (about 450 AA) are not found in the same entry, but are listed under the
genomic scaffolds dumped into NCBI by Celera. Having worked on these domains,
I know they are important for the function of the intact protein. The main
problem (I believe) in identifying the protein is the absence of any
homologous domains, which makes the translated cDNA appear to be a possible
'psuedo gene.' So my next question is given the large number of sequences of
'unknown' function, how does one tell if the protein produced is from a psuedo
gene? Secondly, is there an entry or a compiled list of translated cDNA which
are of 'unknown' function, but have been verified not to be a psuedo gene
product?
.
Thanks for your assistance. You are doing a great job considering how much
data is being generated.
Sincerely,
Thomas L. Selby, Ph. D.
The Scripps Research Institute