On 06/12/2018 09:38 PM, liyaowang2...@163.com wrote:
Hi Edward,
I also have the similaridea, but implement it using the phenix pdb tools.
OK, maybe you are way ahead of me and just want some way to apply the changes
you make in one chain to all the rest.
1. build the model in coot
2. make a dimer
3. phenix pdb tools to roate and translate the dimer based on helical parameters
I see- you already had the helical parameters from the reconstruction
procedure- but with some uncertainty about the location of the axis. If you LSQ
fit one monmer on another, the rot-trans operator includes the information about
the rotation axis, and repeatedly applying it will generate units that spiral
about that axis. (once you've got molecules going half-way around, and
optimized their fit to the density, you can get a more accurate operator by
simultaneously fitting 1 onto 2, 2 onto 3, 3 onto 4 etc.)
Or, assuming the spiral axis is parallel to one axis, say z, get the x,y
coordinates of one atom, the same atom, in every unit. They should fit a circle
in the x-y plane, the center of the circle is the origin, the x,y coordinates
of the axis. So with 3 or more such points you solve
(Xi-Xo)^2 + (Yi-Yo)^2 = R^2 for Xo, Yo, and R. With just two points, if you
know the repeating angle from the helical reconstruction, you can take a line
between those two points as the base of an isosceles triangle and use trig to
see where the third point would need to be to give that angle. That point would
be on the rotation axis.
while I think I made something mistakes, because the result looks like wrong
(distortion) if I create a long filament model (like 1200 A).
Applying rotation-translation should not distort the molecules. If the helical
parameters and axis are not just right it might gradually work its way out of
the density as you go along. Or if just the axis is wrong it might move out of
the density and come back in as you go around.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Yao-Wang Li
School of Life Science, Tsinghua University
Room 255, New Biology Building, Beijing, China, 100084
*From:* Edward A. Berry <mailto:ber...@upstate.edu>
*Date:* 2018-06-11 22:16
*To:* COOT <mailto:COOT@JISCMAIL.AC.UK>
*Subject:* Re: rotation and translation when building model in cryo-em map
Hi Yao-Wang,
If I understand the problem correctly, then if you have built two adjacent
molecules, you can superimpose molecule 1 on molecule 2 to get the operator.
Then apply this operator to molecule 2 to make 3, to moolecule 3 to make 4, and
so on around the helix for as far as you want to go. The operator from two
adjacent molecules won't be very accurate, and errors will add up as you go
unless you rigid-body refine each new monomer in the density before using it to
make the next.
The superposition can be done in coot (LSQ-fit) or from command line with LSQMAN or
(LSQKAB?). For applying the operator to one molecule to generate the next, a quick
survey of coot functions shows "new molecule from symop" but that won't help
you in P1. I didn't see an option to use an arbitrary rot/transl matrix, but that doesn't
mean there isn't one. From the command line you would use pdbset to generate
rotated/translated molecules. Then merge them all into one file.
IFF your helix makes exactly one turn in an integral number of steps like
2,3,4,6, or 12, you could re-define your space group as say P6 with the Z axis
equal to the translation for for complete rotation, Then the copies would be
symm-related and could be generated easily in coot.
On 06/10/2018 11:15 PM, liyaowang2...@163.com wrote:
> Dear all,
>
>
> this is my first thread here, and I am not sure this topic was discussed
or not, please tell me if it was.
>
> Coot I thought is originally for density map of crystal, it can work on
the cryo-em map too. I have a cryo-em map of helical filament and try to build a
model. I can build a model based on the monomer, but I am not sure how to generate
the copies based on the symmetry. I think the key problem is how to determine the
rotation center which based on the map, then the roation center together with the
known helical parameters can do it.
>
> suppose we got the rotation center, how to implement these operations in
Coot?
>
> Certainly, there is a stuipd way which I already did, that is to copy
the refined monomer multiple times and fit them all together, later were refined
using the Phenix.
>
> another way is try to edit the pdb file directly. rotate and tranlate
the monomer, later change the chain ID.
>
> But I want to do it simplely in Coot, could it possible?
>
> Best Wishes,
>
>
>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Yao-Wang Li
> School of Life Science, Tsinghua University
> Room 255, New Biology Building, Beijing, China, 100084
>
>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> To unsubscribe from the COOT list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=COOT&A=1
>
########################################################################
To unsubscribe from the COOT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=COOT&A=1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
To unsubscribe from the COOT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=COOT&A=1
########################################################################
To unsubscribe from the COOT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=COOT&A=1