                    SEGMENT POSITIONING

    Activating the 'Segment Position' item of the 3D  'Edit'
pulldown   menu   allows   one  to  interactively  define  a
collection  of  segments,  each  of  which   may   then   be
repositioned  and  reoriented  as  a  rigid body relative to
predefined sets of axes.  Recalling  that  a  segment  is  a
contiguous  set of bases, the axes sets for each segment are
defined as follows.

    axis set O:  The line connecting the backbone (C1')
                 atoms of the 3' and 5' ends. It is
                 denoted as (S:3'-5').
    axis set 1:  The line connecting the midpoint (MP)
                 of the line (S:3'-5') and the center of
                 mass (COM) of the segment. It is denoted
                 as (S:MP-COM).
    axis set 2:  The normal cartesian X,Y,Z axes
                 positioned at the backbone (C1') atom of
                 the 3' end.  It is denoted as (X,Y,Z:3').
    axis set 3:  The normal cartesian X,Y,Z axes positioned
                 at the backbond (C1') atom of the 5' end.
                 It is denoted as (X,Y,Z:5').
    axis set 4:  The normal cartesian X,Y,Z axes positioned
                 at MP.  It is denoted as (X,Y,Z:MP).
    axis set 5:  The normal cartesian X,Y,Z axes positioned
                 at COM.  It is denoted as (X,Y,Z:COM).

Translation along and rotation about any of these  axes  can
be  performed  to  provide a wide variety of positioning and
orientation   possibilities.    These   possibilities    are
compounded  by  the  feature that for any of the axis sets a
segment can be designated as being closed (includes both  of
its  ends),  open (does not include either end) or semi-open
(includes only one end).

   The purpose of segment location and  orientation  editing
is  for  conveniently  eliminating  any overlapping of bases
that might have  occurred  during  the  2D_to_3D  conversion
process  used  to  obtain  the  initial 3D model from its 2D
counterpart.  It is also used for maniplating the  3D  model
in  a  manner  that will enhance the refinement of specified
tertiary interactions.

   An example of the first use is with regard to the  sample
"pk_mcpheeters"  BPL.  This  is a pseudoknot structure whose
initial 3D  structure  shows  some  potentially  conflicting
overlap  of the single strand (21-25) with the stem shown in
red.  Defining a segment as the bases from 20 through 26 and
then  rotating  it  as  an  open segment by 36 degrees or so
nicely eliminates the  potential  conflict  and  provides  a
favorable  starting  point for refinement, - such as single-
strand refinement in the gobal context mode.

   An example of the second kind of use is  with  regard  to
the  sample  "6tna"  BPL, which corresponds to phenylalanine
transfer RNA.  After doing stem stacking (the first with the
fourth  and  the  second  with the third), there remains the
task of single-strand and  tertiary  bond  refinement.   The
tertiary  bonds  are  colored  in red and it is evident that
they need to be considerably shortened in order  to  achieve
the  classic  3D  form.  The shortening will be accomplished
with global refinement techniques,  but  these  need  to  be
aided  by a preliminary relative movement of the two helices
in order to get around some potential relative  minima  that
will  inhibit  reaching  the  correct  3D form.  A number of
rotation axes can be used to achieve the relative  movement,
among which is the one connecting the base numbers 9 and 49.
A positive rotation of 80 degrees about this  axis  gives  a
favorable starting point for further refinements.

   The kind of segment movement that is employed is somewhat
unreal  because  of the local distortions that it introduces
depending on  whether  the  segment  is  being  regarded  as
closed,  open  or semi-open. These distortions are partially
eliminated by a prerefinement process that is  automatically
invoked  during any of these movements that eliminates large
scale distortions.   Small  scale  distortions   are  easily
corrected  in  subsequent  refinements (single-strand and/or
global), and hence a small price to pay for the facility  of
convenient large scale manipulations.

   Finally, it should be noted that there  is  no  limit  on
segment  size,  large  or  small.  At the large level, whole
branches of bases can be accommodated, while  at  the  small
level an individual nucleotide can be manipulated as the 2nd
base of a three-base sequence designated as open or  as  the
single base of a one-base sequence designated as closed.

                          THE END





