[ssml] Psiblast and tandem repeats

[Richard Harrington]

Dear all,

I'm using psiblast to look at potential homologues of the subunits of a
protein complex.

Two of the subunits have significant hits (e-value < 0.0005) against the
same 3 different proteins.

Subunit A:  ~800aa. Hits on region from ~500-800

hit on CCAAT displacement protein (residues 177-362)
hit on BLACKJACK protein (residues 741-871)
hit on Laminin gamma 1 precursor (residues 1293-1576)

Subunit B:  ~180aa. Hits along whole length of subunit.

hit on CCAAT displacement protein (residues 181-341)
hit on BLACKJACK protein (residues 763-886)
hit on Laminin gamma 1 precursor (residues 1407-1532)


The hits on the BLACKJACK region are in a region of 21 tandem repeats, in
a region of coiled coil potential.
The hits on Laminin are in a region which has no domain annotated, but is
next to a region of consecutive 60aa repeats in a coiled coil structure.
The hits on CCAAT are in a region annotated as CCAAT-like-domain, also
just upstream of a coiled coil domain.

Subunit A and Subunit B do not score significant hits on each other. In an
all against all BLAST of Laminin, Blackjack,and CCAAT, all show
significant similiarity (e-value < 1e-06).

------------------------------------
The questions I have are:

How can I interpret the results of a hit on a region of tandem repeats? Do
the tandem repeats render the score less valid?
Can coiled-coil domains affect a sequence search score?
Can I infer a relationship between the 2 subunits due to their common
potential homologues?
How can I investigate this scenario further?

Many thanks for your time and suggestions,

Richard

_______________________________________________

Richard Harrington,
HUS Graduate Vice-President,
Men's Boat Club Captain,
Homerton College.

PhD student,
Bioinformatics Group,
MRC-Dunn Human Nutrition Unit,
Cambridge University.

e-mail: rah at mrc-dunn.cam.ac.uk
telephone: +44 (0) 1223 252861
_______________________________________________