nucleic acid binding domains

ewan birney birney at molbiol.ox.ac.uk
Tue Mar 7 15:15:41 EST 1995


Newswatcher at sickkids.on.ca (Newswatcher) wrote:
>
> Could anyone suggest a database/software where upon entering the AA
> sequence, possible DNA/RNA binding domains could be found?
> 
> Thanks.
> 
> ajay(ajay at sickkids.on.ca)
> 
> -- 
> Ajay Pandita
> Hospital for Sick Children
> Toronto

I'm afraid to say that for most cases (if not all) you have to do
it yourself. :(

If you are running GCG nearby then you could try profilescan
but this information is (usually) badly out of date.

Prosite, which does have some DNA/RNA binding domains is not ideal
for making these comparisons. Really you need profiles or
hidden markov models and although people (like me also kay
hoffman and philip bucher) are working on libraries of profiles/
HMM's it ain't done yet. I have heard of other places keeping
some, but never found them (any info anyone?)

If you have an amino acid sequence the first thing to do is see
if you can pick up any "obvious" homologies before asking these
more specific questions. This would be using FASTA or BLAST. Then
if you want to ask the question 

"Does my protein have domain xxx" 

the thing to do is build a profile yourself of that domain and 
then try pairwise alignments of your sequence against the profile
probably comparing it to pairwise alignments of a randomised
sequence (from your sequence)...

Not easy...

An outline of this sort of search strategy is given at

http://www.ocms.ox.ac.uk/~birney/wise/search1.html


But, if you like I could test (for you) for the presence of the
RNA-Recognition-Motif (RRM), my "pet" domain. (for what it's worth).


I hope this is of help/interest

ewan

birney at molbiol.ox.ac.uk

 




More information about the Bio-soft mailing list