• ## Outreach

From: Jan Saam (jan.saam_at_rz.hu-berlin.de)
Date: Sun Mar 21 2004 - 05:16:00 CST

Hi Li,

If you want to compute the average structure or the RMSD of the system
throughout your trajectory you should really align all frames first. This
removes rotation and center of mass movements which could lead to distorted
average structure.
But even after the alignment your avg structure will suffer from "unnatural"
atom positions, especially in rotating parts like methyl groups:
The trajectory_smooth script just computes the arithmetic meanof each
coordinate over the specified number of frames. The arithmetic mean position
of methyl H-atoms is somewhere on the rotation axis, leading to an ill
looking conformation of these atoms. But this should be no problem for
studying the structure in general.

Below is a script that does the alignment for you.
Hope this helps,

Jan

#############################################################
# Align all frames of a molecule to the first frame #
#############################################################

proc align_all_frames {sel {mol top}} {

set seltext [\$sel text]
# use frame 0 for the reference
set reference [atomselect \$mol "\$seltext" frame 0]

# the frame being compared
set compare [atomselect \$mol "\$seltext"]
set all [atomselect \$mol all]

set num_steps [molinfo \$mol get numframes]

for {set frame 0} {\$frame < \$num_steps} {incr frame} {
# get the correct frame
\$compare frame \$frame
\$all frame \$frame
# compute the transformation
set trans_mat [measure fit \$compare \$reference]
# do the alignment
\$all move \$trans_mat
# compute the RMSD
set rmsd [measure rmsd \$compare \$reference]
puts "\$frame \$rmsd"
}
display update
}

Am Sonntag, 21. März 2004 09:43 schrieben Sie:
> dear sir :
> thank you for your kind reply .i am sorry for the previous mail.is this
> one ok? i send it again.
>
> i have download the trajectory_smooth script and try to find out the
> average structure of one molecule in waters,but at the same time ,i find
> that the center of the molecule has moved ,since it computes the average
> "postion", does this moving center affect the result?
>
> the same question about rmsd compution.
>
> best wishes!