CCL: how much homology models could be trusted for looking at the dynamic comparisons.

[Arun Kumar Subramanian <aksub007%gmail.com>]

Dear CCL Users,

                       I am trying to see if homology
 model of a GPCR could be trusted in order to compare the low frequency motions
 with other
 homologous proteins for which crystal structures are available. The template
 sequence has 37% similarity and hence I thought of proceeding further because
 most (almost all) of the reported homology models of GPCR are only around this
 value. I would
 guess, if any bias should show up, it should stem out from the fact that
  the fold is simply borrowed from the template and/or the differences in side
 chain orientations. I am thinking of
 reducing this bias via unrestricted, explicitly solvated MD simulations
 using the built homology models or with back bone restrictions, for around 2-3
 ns or upto 5 ns,
 such that the side chain orientations could be treated in a better
 manner. Then, I plan to compare those with the crystal structures on the
  properties of their collective motions or the low frequency movements.
 Any suggestions if this will work out? or any experience in similar
 lines? welcome to share with
 me.

Thanks.

Sincerely,




ArunKumar. S.