Tutorial: Fit Biomolecules into Cryo-EM Maps using MD Simulation (GUI)

Contents

Overview

This tutorial will show you how to fit biomolecule atomic structures into cryo-EM maps using molecular dynamics simulation within the PHENIX graphical user interface (GUI).

For commandline execution, please see the cryo_fit_commandline_tutorial

Theoretical explanation of cryo_fit is here

For installation of cryo_fit, please see the installation notes for cryo_fit

Input files

<initial_model> and <target_map>

Initial Model

Available format: .cif and .pdb

The initial model is a guide or template structure (CIF/mmCIF/pdb) that is close to a target cryo EM map structurally.

You can use either map_to_model or UCSF chimera (Tools -> Volume Data -> Fit in Map)) to prepare the initial model.

Target Map

Available format: .ccp4 and .map (MRC style in binary file) and .sit (Situs style in text file)

Launch cryo_fit

../images/cryo_fit_gui_launch.png

Enter Input Files

Click browse buttons.

(Tutorial input files live in <User_phenix>/modules/cryo_fit/tutorial_input_files)

../images/cryo_fit_gui_enter_input_files.jpg

Specify cryo_fit executable location

Click browse buttons.

(select for exampple, /Users/doonam/bin/cryo_fit/bin)

../images/cryo_fit_gui_specify_bin.jpg

Enter Options

All options can be left blank (cryo_fit will figure out all options automatically).

Run

../images/cryo_fit_gui_run.jpg

It should start something like this (total steps are 1~8).

../images/cryo_fit_gui_it_should_start.jpg

Running time (With 2.7 GHz CPU, macOS)

All default values, it took ~2 minutes. With 10k steps, it took ~9 minues

Result

../images/cryo_fit_gui_result1.png ../images/cryo_fit_gui_result2.png

Output files are in steps/8_cryo_fit folder

A finally fitted atomic model: cryo_fitted.pdb

.gro and .pdb files from the highest 3 cc values: extracted_x_steps_x_ps.gro/pdb (.gro files are for vmd)

CC means 'correlation coefficient between atomic structure and cryo-EM map': steps/8_cryo_fit/cc_record

Output Example

../images/cryo_fit_result.jpg

Limitation

gromacs 4.5.5 seems to not handle H2O (water) heteroatom. cryo_fit will remove water molecules (if any) from the input .cif/.pdb and fit to cryo EM map.

cryo_fit doesn't handle non-canonical "residue"s such as 7C4, BMA, GDP, ILX, NAG, SEP, TRX. The cryo_fit will simply exclude those residues and report what are excluded.

Reference

S. Kirmizialtin, J. Loerke, E. Behrmann, C. MT. Spahn, K. Y Sanbonmatsu, Using Molecular Simulation to Model High-Resolution Cryo-EM Reconstructions, Methods Enzymol., 558, 2015, 497-514

How to generate and record movie?

Generate_record_movie_by_cryo_fit1

List of most useful options

Option Default value Description of inputs and uses
emweight_multiply_by 8 Multiply by this number to the number of atoms for weight for cryo-EM map bias. For example, emweight = (number of atoms in gro file) x (emweight_multiply_by which is 8) The higher the weight, the stronger bias toward EM map rather than MD force field and stereochemistry preserving constraints. If user's map has a better resolution, higher value of emweight_multiply_by is recommended since map has much information. If user's map has have a worse resolution, lower value of emweight_multiply_by is recommended for more likely geometry. If CC (correlation coefficient) needs to be improved faster, higher number of emweight_multiply_by is recommended.
nproc max cores Specify number of cores for minimization and cryo_fit. If it is not specified, or max is chosen, the cryo_fit will try to use most cores automatically (up to 16)
number_of_steps_for_cr yo_fit None This is the initial number of steps for cryo_fit. Eventually, cryo_fit will increase it depending on molecule size and cc trend. For tutorial files, this will be 70,000
number_of_steps_for_mi nimization None Specify number of steps for minimization. If this is left blank, cryo_fit will estimate it depending on molecule size.number of steps for cryo_fit. Enough minimization will prevent "blow-up" during MD simulation later.