· Delarue web servers
· Nomad Flow-chart
· Normal Mode calculation
  · Examples (Movies)
  · Submit a job (from PDB file)
  · Split trajectory (for MR)
  · Generate decoys (MixMod)
  · Elastic Energy (Perturb. Anal.)
· Overlap coefficients
  · Submit a job
  · Include Profit
· X-Ray refinement
  · Standard refinement
  · Screening Mol. Repl. Solns.
· EM refinement
  · Documents/Examples
  · Get Structure Factors
  · Submit a job (no NCS)
  · Submit job with NCS
· Docking refinement
  · Submit a job
· Force field methods
  · Energy minimization
  · Gromacs NMA
· References
· Homes/Links
· Job queue status


We have implemented the following new features in our program:
-Incoporate Non-Crystallographic Symmetry in the refinement of X-Ray models in EM-maps
-Add an R-free (CC-free) criterion to evaluate overfitting
-Provide for an example (test-case) with GroEL (7-fold symmetry)
-Make sure the calculation can be done on an entire virus particle

General Remarks

Although Normal Modes Refinement can function as Rigid-Body refinement by using the first 6 modes, which are overall rotation and translation modes, this will not work properly if there is a large rotational component, as Normal Modes eigenvectors are essentially linear approximations of the overall rotations.

Example 1 (ATPase)

Using the same experimental map as Hinsen and coll. (kindly provided by the authors), the phased correlation coefficient of the initial model increased from 0.383 to 0.568 using 21 (15+6) modes and data between 100 and 10 Angstrom resolution.
The geometry of the model remained acceptable during refinement.
As a control, the final model of Hinsen and coll., obtained in a completely independent way, had a phased correlation coefficient of 0.547.
At the same time, the corr. coeff. on Amplitudes of Structure Factors increased only from 0.71 to 0.72, whereas the final model of Hinsen et al. had a correlation coefficient of 0.77.
Beware: the output of the program also displays a list of violations of CA-CA distances (normally 3.8 Angstroms) in the refined model. If there are too many of these violations, you should consider the model as doubtful and try a different number of modes.

Example 2 (GroEL)

Using the cryo-EM map (displaying 7-fold non-crystallographic symmetry) made available at an EMBO workshop that recently took place in Gif/Yvette and organized by J. Navaza, we tested the ability of our program to refine the unbound form of GroEL into the EM-map of the ATP-bound form.
The map was inverted to get structure factors between 25 and 150 Angstrom resolution (1225 Refl., out of which 10% were left out of the refinement process to calculate a CC-free agreement factor).
Normal modes were calculated on the monomer and the program made full use of the 7-fold Non-crystallographic symmetry. Modes 1-21 were allowed to adjust their amplitudes, resulting in an increase of the phased correlation coefficient from 0.55 (CC-free=0.64) to 0.83 (CC-free=0.91). The calculation took about 30 mn cpu.
The rmsd between the initial and the refined model was about 11. Angstrom, indicative of large rearrangements that are probably beyond the reach of the elastic model. However, good stereochemistry (i.e. plausible CA-CA distances) was observed in the refined model.
It is likely that better strategies should be used for this kind of very large rearrangements, such as applying only part of the corrections to the model indicated by the program, followed by refinement of stereochemistry and redetermination of the normal modes before applying another round of the refinement.
As a comparison, a model refined in the same density map by the program NORMA of K. Suhre (CNRS, MArseille) achieved a better correlation coefficient (0.91), using a more careful strategy and using URO (J. Navaza, CNRS, Gif/Yvette) to take into account NCS and rigid-body refinement.

Example 3 (viral particle)

We checked that the cpu time required to refine a full T=3 virus structure is reasonable (less than an hour) and gives satisfactory results.

Related publications

The article of K. Hinsen and coll. (2005) describes a similar method to perform normal mode refinement of a model in an experimental EM map; the difference with our work is that they work in real space (see their Pubmed article) whereas we work in reciprocal space.


Here are a few hints on how to get structure factors from your experimental map:

1. First you must have a complete asymmetric unit ot the map, preferentially a CNS map. Here is the format used by CNS and some code to read and write such maps: map_cns_format

2. You may convert and invert it using
a/ Rave and mapman from G. Kleywegt USF and
b/ FFT from CCP4.
Here is a typical command file to do so: sfall

3. If you don't want to bother with this, you can try the "Get Str. Fact." option on the left. But read the documentation first!

Related links

For maps handling, structure factor calculations... see CCP4 , CNS and Uppsala Software Factory (USF) web sites.

For other tools fitting models into low-resolution maps, based on the same principles:

See NMFF web site from Fl. Tama and C. Brooks.

See NORMA web site from K. Suhre.

  Marc Delarue http://lorentz.dynstr.pasteur.fr
Page last modified 12:32 January 10, 2007.