Flex-EM

A method for protein structure fitting and refinement guided by cryoEM density.

Reference: M. Topf, K. Lasker, B. Webb, H. Wolfson, W. Chiu, A. Sali. Structure 16, 295-307, 2008

The method includes a rigid fitting stage followed by a refinement stage. Rigid fitting is performed with the Mod-EM method (using the Monte-Carlo protocol). Refinement is performed using the scripts below. Two methods are available: Conjugate Gradients (CG) and Molecular Dynamics (MD). The scripts run with MODELLER 9v2 or later.

Download the following scripts:

• CG.py
• MD.py
• rigid.py
• flex-em.py
• example-rigid.txt

Edit the following files:

1. flex-em.py:

   Set the input parameters (the EM map parameters and the name of the input structure) and define the type of optimization (CG or MD). If you have a multi-domain protein and you are not sure about the approximate position of a domain, it is recommended to (a) run the CG minimization multiple times (e.g., "num_of_runs = 20" in flex-em.py) and (b) cluster the different "fits".

2. rigid.txt:

   Define the rigid bodies. The format of this file is the following (see example-rigid.txt for an example):
   • A comment line describing what kind of a rigid body : "domain", "helix", "beta", "loop", followed by lines describing the rigid bodies : each line describes one rigid body by specifying the initial and final residue of each of the segments in that rigid body.
   • For example: 36 50 55 100 means that residues 36-50 and 55-100 will be included in the rigid body. (Note that before assigning the rigid bodies the structure has to be renumbered from 1 because MODELLER is going to renumber it during the optimization).
   • If you add "nr" after the rigid body the position of this rigid body will not be randomized prior to running of the CG minimization.

Run Flex-EM using the following command:

mod9v2 flex-em.py 

Flex-EM can be applied to atomic structures and models. A tutorial can be found as part of a tutorial on combining comparative modeliing and cryoEM density fitting.

Benchamrk

Flex-EM benchmark is available for download.

Notes:

• To minimize bias, the benchmark density maps were not produced with Mod-EM, but with pdb2vol in SITUS (Wriggers, et al. 1999), which uses a different Gaussian smoothing technique.
• Cα RMSDs were calculated for the benchmark using the superpose method in MODELLER 9v0.

For further information, please contact Maya Topf at