[ssml] Re: Sequence defined domains?

[Dan Bolser]

Cheers,

Here is the reply I got from Robert Finn at Pfam...

> Do you have any examples from Pfam?

Yes, there are several examples.  Cases where we know the presence of a
"nested" domain are indicated in the description files with NE (nested
pfam identifier)and NL (nested domain location). IMPDH domain is come
examples contains two CBS domains.  The multisubunit RNA polyermase
largest subunits (Rpb1 and Rpb2) also contain nested domains.

These are special situation though, as pfam does not normally allow
overlaps between domain.  This is a relatively new feature for Pfam and we
are currently adding more examples of such domains.

However, searching genomes with our HMMs does flag up cases where a domain
contains an inserted domain.

Hope this helps!

Cheers,

Rob

====================================================================

Robert Finn
The Wellcome Sanger Institute
Wellcome Trust Genome Campus
Hinxton
Cambs
CB10 1SA
Tel:+44 (0)1223 495330
====================================================================



<quote who="Kevin Karplus">
> I don't know whether Pfam has any HMMs for split domains, but
> Superfamily certainly does, because Julian Gough and I had several discussions
> about the best way to represent them with SAM HMMs.
> It turned out that the usual SAM-T99 or SAM-T2k script does a pretty good job if
> this the inserted domain is put in lower-case and the
> surrounding domain is put in upper-case in the initial seed.
>
> Kevin Karplus
> _______________________________________________
> ssml-general mailing list
> ssml-general at bioinformatics.org
> https://bioinformatics.org/mailman/listinfo/ssml-general