On Mon, 2012-11-12 at 13:47 -0500, Ed Pozharski wrote:
> Does anyone know of a tool that would generate a protein molecule
> backbone from a set of phi/psi angles?
>
For the record.
Thanks to all who responded. Here is what I found out:
1. MOLEMAN works best. The most cumbersome part is to transfer phi/psi
from one structure to another. If anyone is ever interested, I have a
bash script that does it in automated manner, all you need to do is to
provide the list of residues to be corrected.
2. Other options include TINKER and fortran code by Edward Berry. I
couldn't use the second one because it does not do omegas and I have
some cis-prolines in the model. TINKER is an awesome MD package from
what I can tell. I abandoned it for moleman after concluding that it
takes integers values, but I didn't investigate if that is actually true
and if it is really a big problem (see below).
3. On the scientific side, it appears that updating phi/psi angles is
not enough. Basically, even if I transfer all the angles from one model
to another, models match well only in the N-terminal domain. The
C-terminal domain which in this case undergoes about 45 degrees
closure-type rotation is very much messed up. Perhaps the conclusion is
that changes other than phi/psi contribute significantly. Which is, on
the second thought, not that surprising.
Cheers,
Ed.
--
After much deep and profound brain things inside my head,
I have decided to thank you for bringing peace to our home.
Julian, King of Lemurs
