The Coot User Manual

Table of Contents

Coot User Manual

 — The Detailed Node Listing —

Introduction

Mousing and Keyboarding

General Features

Scripting

Python


Backups and Undo


Coordinate-Related Features

Modelling and Building

Regularization and Real Space Refinement 

Map-Related Features 

Validation 

Hints and Usage Tips

Other Programs

1 Introduction

This document is the Coot User Manual, giving an overview of the interactive features. Other documentation includes the Coot Reference Manual and the Coot Tutorial. These documents should be distributed with the source code.

1.1 Citing Coot and Friends

If have found this software to be useful, you are requested (if appropriate) to cite:

"Features and Development of Coot" P Emsley, B Lohkamp, W Scott, and K Cowtan Acta Cryst. (2010). D66, 486-501 Acta Crystallographica Section D-Biological Crystallography 66: 486-501

The reference for the REFMAC5 Dictionary is:

REFMAC5 dictionary: "Organization of Prior Chemical Knowledge and Guidelines for its Use" Vagin AA, Steiner RA, Lebedev AA, Potterton L, McNicholas S Long F, Murshudov GN Acta Crystallographica Section D-Biological Crystallography 60: 2184-2195 Part 12 Sp. Iss. 1 DEC 2004"

If using "SSM Superposition", please cite:

"Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions" Krissinel E, Henrick K Acta Crystallographica Section D-Biological Crystallography 60: 2256-2268 Part 12 Sp. Iss. 1 DEC 2004

The reference for the the Electron Density Server is:

GJ Kleywegt, MR Harris, JY Zou, TC Taylor, A Wählby, TA Jones (2004), "The Uppsala Electron-Density Server", Acta Crystallographica Section D-Biological Crystallography 60, 2240-2249.

Please also cite the primary literature for the received structures.

1.2 What is Coot?

Coot is a molecular graphics application. Its primary focus is crystallographic macromolecular model-building and manipulation rather than representation i.e. more like Frodo than Rasmol. Having said that, Coot can work with small molecule (SHELXL) and electron microscopy data, be used for homology modelling, make passably pretty pictures and display NMR structures.

Coot is Free Software. You can give it away. If you don’t like the way it behaves, you can fix it yourself.

1.3 What Coot is Not

Coot is not:

• CCP4’s official Molecular Graphics program ¹
• a program to do refinement ²
• a protein crystallographic suite³.

1.4 Hardware Requirements

The code is designed to be portable to any Unix-like operating system. Coot certainly runs on RedHat Linux of various sorts, Fedora, Ubuntu, Debian, SuSe Linux and MacOS X. There is also a Window port (called WinCoot).

If you want to port to some other operating system, you are welcome ⁴. Note that your task will be eased by using GNU GCC to compile the programs components.

1.4.1 Mouse

Coot works best with a 3-button mouse and works better if it has a scroll-wheel too (see Chapter 2 for more details) ⁵.

1.5 Environment Variables

Coot responds to several environment variables that modify its behaviour.

• COOT_STANDARD_RESIDUES The filename of the pdb file containing the standard amino acid residues in “standard conformation” ⁶
• COOT_SCHEME_DIR The directory containing standard (part of the distribution) scheme files
• COOT_SCHEME_EXTRAS_DIR A ’:’-separated list of directories containing bespoke scheme files. This variable is not set by default. If you set it, Coot will test each ’:’-separated string that it points to a directory, and if it does, Coot will load all the .scm files in that directory.
• COOT_PYTHON_EXTRAS_DIR A ’:’-separated list of directories containing bespoke python files. This variable is not set by default. If you set it, Coot will test each ’:’-separated string that it points to a directory, and if it does, Coot will load all the .py files in that directory.
• COOT_REF_STRUCTS The directory containing a set of high resolution pdb files used as reference structures to build backbone atoms from C\alpha positions
• COOT_REF_SEC_STRUCTS The directory containing a set of high-quality structures to be used as templates for fitting beta strands. If this is not set, then the directory COOT_REF_SEC_STRUCTS will be used to find the reference pdb files.
• COOT_REFMAC_LIB_DIR Refmac’s CIF directory containing the monomers and link descriptions. In the future this may simply be the same directory in which refmac looks to find the library dictionary.
• COOT_SBASE_DIR The directory to find the SBASE dictionary (often comes with CCP4).
• COOT_RESOURCES_DIR The directory that contains the splash screen image and the GTK+ application resources.
• COOT_BACKUP_DIR The directory to which backup are written (if it exists as a directory). If it is not, then backups are written to the current directory (the directory in which coot was started).

And of course extension language environment variables are used too:

• PYTHONPATH (for python modules)
• GUILE_LOAD_PATH (for guile modules)

Normally, these environment variables will be set correctly in the coot shell script.

1.6 Command Line Arguments

Rather that using the GUI to read in information, you can use the following command line arguments:

• --c cmd to run a command cmd on start up
• --script filename to run a script on start up (but see Section Scripting)
• --no-state-script don’t run the 0-coot.state.scm script on start up. Don’t save a state script on exit either.
• --pdb filename for pdb/coordinates file
• --coords filename for SHELX .ins/.res and CIF files
• --data filename for mtz, phs or mmCIF data file
• --auto filename for auto-reading mtz files (mtz file has the default labels FWT, PHWT)
• --map filename for a map (currently CCP4-format only)
• --dictionary filename read in a cif monomer dictionary
• --help print command line options
• --stereo start up in hardware stereo mode
• --version print the version of coot and exit
• --code accession-code on starting Coot, get the pdb file and mtz file (if it exists) from the EDS
• --no-guano don’t leave “Coot droppings” i.e. don’t write state and history files on exit
• --side-by-side start in side-by-side stereo mode
• --update-self command-line mode to update the coot to the latest pre-release on the server
• --python an argument with no parameters - used to tell Coot that the -c arguments should be process as python (rather than as scheme).
• --small-screen start with smaller icons and font to fit on small screen displays
• --zalman-stereo start in Zalman stereo mode

So, for example, one might use:

• coot --pdb post-refinement.pdb --auto refmac-2.mtz --dictionary lig.cif

1.7 Web Page

Coot has a web page:

• http://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/

There you can read more about the CCP4 molecular graphics project in general and other projects which are important for Coot ⁷.

1.8 Crash

Coot might crash on you - it shouldn’t.

Whenever Coot manipulates the model, it saves a backup pdb file. There are backup files in the directory coot-backup ⁸. You can recover the session (until the last edit) by reading in the pdb file that you started with last time and then use File -> Recover Session….

I would like to know about coot crashing ⁹ so that I can fix it as soon as possible. If you want your problem fixed, this involves some work on your part sadly.

First please make sure that you are using the most recent version of coot. I will often need to know as much as possible about what you did to cause the bug. If you can reproduce the bug and send me the files that are needed to cause it, I can almost certainly fix it ¹⁰ - especially if you use the debugger (gdb) and send a backtrace too¹¹. Note that you may have to source the contents of bin/coot so that the libraries are can be found when the executable dynamically links.

2 Mousing and Keyboarding

How do we move around and select things?

Left-mouse Drag

Rotate view

Ctrl Left-Mouse Drag

Translates view

Shift Left-Mouse

Label Atom

Right-Mouse Drag

Zoom in and out

Ctrl Shift Right-Mouse Drag

Rotate View around Screen Z axis

Middle-mouse

Centre on atom

Scroll-wheel Forward

Increase map contour level

Scroll-wheel Backward

Decrease map contour level

See also Chapter Hints and Usage Tips for more help.

2.1 Next Residue

``Space''

Next Residue

``Shift'' ``Space''

Previous Residue

See also “Recentring View” (Section Recentring View).

2.2 Keyboard Contouring

Use + or - on the keyboard if you don’t have a scroll-wheel.

2.3 Mouse Z Translation and Clipping

Here we can change the clipping and Translate in Screen Z

Ctrl Right-Mouse Drag Up/Down

changes the slab (clipping planes)

Ctrl Right-Mouse Drag Left/Right

translates the view in screen Z

2.4 Keyboard Translation

Keypad 3

Push View (+Z translation)

Keypad .

Pull View (-Z translation)

2.5 Keyboard Zoom and Clip

N

Zoom out

M

Zoom in

D

Slim clip

F

Fatten clip

2.6 Scrollwheel

When there is no map, using the scroll-wheel has no effect. If there is exactly one map displayed, the scroll-wheel will change the contour level of that map. If there are two or more maps, the map for which the contour level is changed can be set using either HID -> Scrollwheel -> Attach scroll-wheel to which map? and selecting a map number or clicking the "Scroll" radio button for the map in the Display Manager.

You can turn off the map contour level changing by the scroll wheel using:

(set-scroll-by-wheel-mouse 0)

(the default is 1 [on]).

2.7 Selecting Atoms

Several Coot functions require the selecting of atoms to specify a residue range (for example: Regularize, Refine (Section Regularization and Real Space Refinement) or Rigid Body Fit Zone (Section Rigid Body Refinement)). Select atoms with the Left-mouse. See also Picking (Section sec_picking).

Use the scripting function (quanta-buttons) to make the mouse functions more like other molecular graphics programs to which you may be more accustomed ¹².

2.8 Virtual Trackball

You may not completely like the way the molecule is moved by the mouse movement ¹³. To change this, try: HID -> Virtual Trackball -> Flat. To do this from the scripting interface: (vt-surface 1) ¹⁴.

If you do want screen-z rotation screen-z rotation, you can either use Shift Right-Mouse Drag or set the Virtual Trackball to Spherical Surface mode and move the mouse along the bottom edge of the screen.

2.9 More on Zooming

The function (quanta-like-zoom) adds the ability to zoom the view using just Shift + Mouse movement ¹⁵.

There is also a Zoom slider (Draw -> Zoom) for those without a right-mouse button.

3 General Features

The map-fitting and model-building tools can be accessed by using Calculate -> Model/Fit/Refine…. Many functions have tooltips ¹⁶ describing the particular features and are documented in Chapter Modelling and Building.

F5:

posts the Model/Fit/Refine dialog

F6:

posts the Go To Atom Window

F7:

posts the Display Control Window

3.1 Version number

The version number of Coot can be found at the top of the “About” window (Help -> About).

This will return the version of coot:

$ coot --version

There is also a script function to return the version of coot:

(coot-version)

3.2 Antialiasing

The built-in antialiasing (for what it’s worth) can be enabled using:

(set-do-anti-aliasing 1)

The default is 0 (off).

This can also be activated using Edit Preferences -> Others -> Antialiasing -> Yes.

If you have an nVidia graphics card, external antialiasing can be actived setting the environment variable __GL_FSAA_MODE. For me a setting of 5 works nicely and gives a better image than using Coot’s built-in antialiasing.

Also for nVidia graphics card users, there is the application nvidia-settings:

Antialiasing Setting -> Override Application Settings and slide the slider to the right. On restarting Coot, it should be in antialias mode ¹⁷.

3.3 Molecule Number

Coot is based on the concept of molecules. Maps and coordinates are different representations of molecules. The access to the molecule is via the molecule number. It is often important therefore to know the molecule number of a particular molecule.

The Molecule Number of a molecule can be found by clicking on an atom of that molecule (if it has coordinates of course). The first number in brackets in the resulting text in the status bar and console is the Molecule Number. The Molecule Number can also be found in Display Control window (Section Display Manager). It is also displayed on the left-hand side of the molecule name in the option menus of the “Save Coordinates” and “Go To Atom” windows.

3.4 Display Issues

The “graphics” window is drawn using OpenGL. It is considerably smoother (i.e. more frames/sec) when using a 3D accelerated X server.

The view is orthographic (i.e. the back is the same size as the front). The default clipping is about right for viewing coordinate data, but is often a little too “thick” for viewing electron density. It is easily changed (see Section Clipping Manipulation).

Depth-cueing is linear and fixed on.

The graphics window can be resized, but it has a minimum size of 400x400 pixels.

3.4.1 Stereo

Hardware Stereo is an option for Coot (Draw -> Stereo… -> Hardware Stereo -> OK), side-by-side stereo is not an option.

The angle between the stereo pairs (the stereo separation) can be changed to suit your personal tastes using:

(set-hardware-stereo-angle-factor angle-factor)

where angle-factor would typically be between 1.0 and 2.0

3.4.2 Pick Cursor

When asked to pick a residue or atom, the cursor changes from the normal arrow shape to a "pick" cursor. Sometimes it is difficult to see the default pick cursor, so you can change it using the function

(set-pick-cursor-index i)

where i is an integer less than 256. The cursors can be viewed using an external X program:

xfd -fn cursor

3.4.3 Origin Marker

A yellow box called the “origin marker” marks the origin. It can be removed using:

(set-show-origin-marker 0)

Its state can be queried like this:

(show-origin-marker-state)

which returns an number (0 if it is not displayed, 1 if it is).

3.5 Screenshot

A simple screenshot (image dump) can be made using Draw -> Screenshot -> Simple…. Note that in side by side stereo mode you only get the left-hand image.

3.6 Raster3D output

Output suitable for use by Raster3D’s “render” can be generated using the scripting function

(raster3d file-name)

where file-name is such as "test.r3d" ¹⁸.

There is a keyboard key to generate this file, run “render” and display the image: Function key F8.

You can also use the function

(render-image)

which will create a file coot.r3d, from which “render” produces coot.png. This png file is displayed using ImageMagick’s display program (by default). Use something like:

(set! coot-png-display-program "gqview")

to change that to different display program ("gqview" in this case).

(set! coot-png-display-program "open")

would use Preview (by default) on Macintosh.

To change the widths of the bonds and density “lines” use (for example):

(set-raster3d-bond-thickness 0.1)

and

(set-raster3d-density-thickness 0.01)

Similarly for bones:

(set-raster3d-bone-thickness 0.05)

To turn off the representations of the atoms (spheres):

(set-renderer-show-atoms 0)

3.7 Display Manager

This is also known as “Map and molecule (coordinates) display control”. Here you can select which maps and molecules you can see and how they are drawn ¹⁹. The “Display” and “Active” are toggle buttons, either depressed (active) or undepressed (inactive). The “Display” buttons control whether a molecule (or map) is drawn and the “Active” button controls if the molecule is clickable ²⁰ (i.e. if the molecule’s atoms can be labeled).

The "Scroll" radio buttons sets which map is has its contour level changed by scrolling the mouse scroll wheel.

By default, the path names of the files are not displayed in the Display Manager. To turn them on:

(set-show-paths-in-display-manager 1)

If you pull across the horizontal scrollbar in a Molecule view, you will see the “Render as” menu. You can use this to change between normal “Bonds (Colour by Atom)”,“Bonds (Colour by Chain)” and “C\alpha” representation There is also available “No Waters” and “C\alpha + ligands” representations.

3.8 The Modelling Toolbar

You might not want to have the right-hand-side vertical toolbar that contains icons for some modelling operations ²¹ displayed:

(hide-modelling-toolbar)

to bring it back again:

(show-modelling-toolbar)

3.9 The file selector

3.9.1 File-name Filtering

The “Filter” button in the fileselection filters the filenames according to extension. For coordinates files the extensions are “.pdb” “.brk” “.mmcif” and others. For data: “.mtz”, “.hkl”, “.phs”, “.cif” and for (CCP4) maps “.ext”, “.msk” and “.map”. If you want to add to the extensions, the following functions are available:

• (add-coordinates-glob-extension extension)
• (add-data-glob-extension extension)
• (add-map-glob-extension extension)
• (add-dictionary-glob-extension extension)

where extension is something like: ".mycif".

If you want the fileselection to be filtered without having to use the "Filter" button, use the scripting function

(set-filter-fileselection-filenames 1)

3.9.2 Filename Sorting

If you like your files initially sorted by date (rather than lexicographically, which is the default) use:

(set-sticky-sort-by-date)

3.9.3 Save Coordinates Directory

Some people prefer that the fileselection for saving coordinates starts in the original directory (rather than the directory from which they last imported coordinates). This option is for them:

(set-save-coordinates-in-original-directory 1)

3.10 Scripting

There is an compile-time option of adding a script interpreter. Currently the options are python and guile. It seems possible that in future you will be able to use both in the same executable. The binary distribution of Coot are linked with guile, others with python.

Hundreds of commands are made available for use in scripting by using SWIG, some of which are documented here. Other functions documented less well, but descriptions for them can be found at the end of this manual.

Commands described throughout this manual (such as (vt-surface 1)) can be evaluated directly by Coot by using the “Scripting Window” (Calculate -> Scripting…). Note that you type the commands in the upper entry widget and the command gets echoed (in red) and the return value and any output is displayed in the text widget lower (green). The typed command should be terminated with a carriage return ²². Files ²³ can be evaluated (executed) using Calculate -> Run Script….

Note that in scheme (the usual scripting language of Coot), the parentheses are important.

To execute a script file from the command line use the --script filename arguments (except when also using the command line argument --no-graphics, in which case you should use -s filename).

After you have used the scripting window, you may have noticed that you can no longer kill Coot by using Ctrl-C in the console. To recover this ability:

(exit)

in the scripting window.

3.10.1 Python

 * Python Commands

Coot has an (optional) embedded python interpreter. Thus the full power of python is available to you. Coot will look for an initialization script ($HOME/.coot.py) and will execute it if found. This file should contain python commands that set your personal preferences.

3.10.1.1 Python Commands

The scripting functions described in this manual are formatted suitable for use with guile, i.e.:

(function arg1 arg2…)

If you are using Python instead: the format needs to be changed to:

function(arg1,arg2…)

Note that dashes in guile function names become underscores for python, so that (for example) (raster-screen-shot) becomes raster_screen_shot().

3.10.2 Scheme

The scheme interpreter is made available by embedding guile. The initialization script used by this interpreter is $HOME/.coot. This file should contain scheme commands that set your personal preferences.

3.10.3 Coot State

The “state” of Coot is saved on Exit and written to a file called 0-coot.state.scm (scheme) 0-coot.state.py (python). This state file contains information about the screen centre, the clipping, colour map rotation size, the symmetry radius, and other molecule related parameters such as filename, column labels, coordinate filename etc..

Use Calculate -> Run Script… to use this file to re-create the loaded maps and models that you had when you finished using Coot ²⁴ last time. A state file can be saved at any time using (save-state) which saves to file 0-coot.state.scm or (save-state-filename "thing.scm") which saves to file thing.scm.

When Coot starts it can optionally run the commands in 0-coot.state.scm.

Use (set-run-state-file-status i) to change the behaviour: i is 0 to never run this state file at startup, i is 1 to get a dialog option (this is the default) and i is 2 to run the commands without question.

3.10.4 Key Binding

“Power users” of Coot might like to write their own functions and bind that function to a keyboard key. How do they do that?

By using the add-key-binding function:

(add-key-binding function-name key function)

where key is a quoted string (note that upper case and lower case keys are distinguished - activate get upper case key binding you need to chord the shift key ²⁵).

for example:

(add-key-binding "Refine Active Residue with Auto-accept" "x" refine-active-residue)

Have a look at the key bindings section on the Coot wiki for several more examples.

3.10.5 User-Defined Functions

“Power users” of Coot might also like to write their own functions that occur after picking an atom (or a number of atoms)

(user-defined-click n_clicks udfunc)

define a function func which runs after the user has made n_clicked atom picks. func is called with a list of atom specifiers - the first member of which is the molecule number.

3.11 Backups and Undo

 * Redo::
 * Restoring from Backup::

By default, each time a modification is made to a model, the old coordinates are written out ²⁶. The backups are kept in a backup directory and are tagged with the date and the history number (lower numbers are more ancient ²⁷). The “Undo” function discards the current molecule and loads itself from the most recent backup coordinates. Thus you do not have to remember to “Save Changes” - coot will do it for you ²⁸.

If you have made changes to more than one molecule, Coot will pop-up a dialog box in which you should set the “Undo Molecule” i.e. the molecule to which the Undo operations will apply. Further Undo operations will continue to apply to this molecule until there are none left. If another Undo is requested Coot checks to see if there are other molecules that can be undone, if there is exactly one, then that molecule becomes the “Undo Molecule”, if there are more than one, then another Undo selection dialog will be displayed.

You can set the undo molecule using the scripting function:

(set-undo-molecule imol)

If for reasons of strange system²⁹ requirements you want to remove the path components of the backup file name you can do so using:

(set-unpathed-backup-file-names 1)

3.11.1 Redo

The “undone” modifications can be re-done using this button. This is not available immediately after a modification ³⁰.

3.11.2 Restoring from Backup

There may be certain circumstances ³¹ in which you wish to restore from a backup but can’t get it by the “Undo” mechanism described above. In that case, start coot as normal and then open the (typically most recent) coordinates file in the directory coot-backup (or the directory pointed to the environment variable COOT_BACKUP_DIR if it was set) . This file should contain your most recent edits. In such a case, it is sensible for neatness purposes to immediately save the coordinates (probably to the current directory) so that you are not modifying a file in the backup directory.

See also Section Crash.

3.12 View Matrix

It is sometimes useful to use this to orient the view and export this orientation to other programs. The orientation matrix of the view can be displayed (in the console) using:

(view-matrix)

Also, the internal representation of the view can be returned and set using:

(view-quaternion) to return a 4-element list

(set-view-quaternion i j k l) which sets the view quaternion.

So the usage of these functions would be something like:

(let ((v (view-quaternion)))
   ;; manipulate v here, maybe
   (apply set-view-quaternion v))

3.13 Space Group and Symmetry

Occasionally you may want to know the space group of a particular molecule. Interactively (for maps) you can see it using the Map Properties button in the Molecule Display Control dialog.

There is a scripting interface function that returns the space group for a given molecule ³²:

(show-spacegroup imol)

You can force a space group onto a molecule using the following:

(set-space-group imol space-group)

where space-group is one of the standard CCP4 space group names (e.g. "P 21 21 21").

To show the symmetry operators of a particular molecule use: (get-symmetry imol) which will return a list of strings.

Sometimes molecular replacement solutions (for example) create models with chains non-optimally placed relative to each other - a symmetry-related copy would be more apealling (but would be equivalent, crystalographically). For example, to move the B chain to a symmetry-related position:

Centre on an atom in the symmetry-related B chain (where you want the B chain to be)

Extensions -> Modelling -> Symm Shift Reference Chain Here.

3.14 Recentring View

• Use Control + left-mouse to drag around the view
• or
• middle-mouse over an atom. In this case, you will often see “slide-recentring”, the graphics smoothly changes between the current centre and the newly selected centre.
• or
• Use Draw -> Go To Atom… to select an atom using the keyboard. Note that you can subsequently use “Space” in the “graphics” window (OpenGL canvas) to recentre on the next C\alpha.
• or
• To centre on an arbitrary position (x,y,z), use the scripting function (set-rotation-centre x y z).
• or
• Use the keyboard: [Ctrl G] then key in a residue number and (optionally) a chainid and press Return

If you don’t want smooth recentring (sliding) Edit -> Preferences -> Smooth Recentring -> Off. You can also use this dialog to speed it up a bit (by decreasing the number of steps instead of turning it off).

3.15 Views

Coot has a views interface (you might call them ”scenes“) that define a particular orientation, zoom and view centre. Coot and linearly interpolate between the views. The animation play back speed can be set with the ”Views Play Speed“ menu item - default is a speed of 10.

The views interface can be found under the Extensions menu item.

3.16 Clipping Manipulation

The clipping planes (a.k.a. “slab” ) can be adjusted using Edit -> Clipping and adjusting the slider. There is only one parameter to change and it affects both the front and the back clipping planes ³³. The clipping can also be changed using keyboard “D” and “F”.

It can also be changed with Ctrl + Right-mouse drag up and down. Likewise the screen-Z can be changed with Ctrl + Right-mouse left and right ³⁴.

One can “push” and “pull” the view in the screen-Z direction using keypad 3 and keypad “.” (see Section Keyboard Z Translation).

3.17 Background colour

The background colour can be set either using a GUI dialog (Edit$ -> Background Colour) or the function (set-background-colour 0.00 0.00 0.00), where the arguments are 3 numbers between 0.0 and 1.0, which respectively represent the red, green and blue components of the background colour. The default is (0.0, 0.0, 0.0) (black).

3.18 Unit Cell

If coordinates have symmetry available then unit cells can be drawn for molecules (Draw -> Cell & Symmetry -> Show Unit Cell?).

3.19 Rotation Centre Pointer

There is a pink pointer at the centre of the screen that marks the rotation centre. The size of the pointer can be changed using Edit -> Pink Pointer Size… or using scripting commands: (set-rotation-centre-size 0.3).

3.20 Orientation Axes

The green axes showing the orientation of the molecule are displayed by default. To remove them use the scripting function;

(set-draw-axes 0)

3.21 Pointer Distances

The Rotation Centre Pointer is sometimes called simply “Pointer”. One can find distances to the pointer from any active set of atoms using “Pointer Distances” (under Measures). If you move the Pointer (e.g. by centering on an atom) and want to update the distances to it, you have to toggle off and on the “Show Pointer Distances” on the Pointer Distances dialog.

3.22 Crosshairs

Crosshairs can be drawn at the centre of the screen, using either the C key³⁵ in graphics window or Draw -> Crosshairs…. The ticks are at 1.54Å, 2.7Å and 3.8Å.

3.23 3D Annotations

Positions in 3D space can be annotated with 3D text. The mechanism to do this can be found under Extensions -> Representations -> 3D Annotations. 3D Annotations can be saved to and loaded from a file.

3.24 Frame Rate

Sometimes, you might as yourself “how fast is the computer?” ³⁶. Using Calculate -> Frames/Sec you can see how fast the molecule is rotating, giving an indication of graphics performance. It is often better to use a map that is more realistic and stop the picture whizzing round. The output is written to the status bar and the console, you need to give it a few seconds to “settle down”. It is best not to have other widgets overlaying the GL canvas as you do this.

The contouring elapsed time ³⁷ gives an indication of CPU performance.

3.25 Program Output

Due to its “in development” nature (at the moment), Coot produces a lot of “console” ³⁸ output - much of it debugging or “informational”. This will go away in due course. You are advised to run Coot so that you can see the console and the graphics window at the same time, since feedback from atom clicking (for example) is often written there rather than displayed in the graphics window.

• Output that starts “ERROR...” is a programming problem (and ideally, you should never see it).
• Output that starts “WARNING...” means that something probably unintended happened due to the unexpected nature of your input or file(s).
• Output that starts “DEBUG...” has (obviously enough) been added to aid debugging. Most of them should have been cleaned up before release, but as Coot is constantly being developed, a few may slip through. Just ignore them.

4 Coordinate-Related Features

4.1 Reading coordinates

The format of coordinates that can be read by coot is either PDB or mmCIF. To read coordinates, choose File -> Read Coordinates from the menu-bar. Immediately after the coordinates have been read, the view is (by default) recentred to the centre of this new molecule and the molecule is displayed. The recentring of the view after the coordinates have been read can be turned off by unclicking the "Recentre?" radio-button.

To disable the recentring of the view on reading a coordinates file via scripting, use: (set-recentre-on-read-pdb 0). However, when reading a coordinates file from a script it is just as good (if not better) to use (handle-read-draw-molecule-with-recentre filename 0) - the additional 0 means “don’t recentre”. And that affects just the reading of filename and not subsequent files.

By default coot does not allow reading coordinates with duplicated sequence numbers. To enable the reading of files with duplicated sequence numbers use the function:

(allow-duplicate-sequence-numbers)

Coot can read MDL mol files.

4.1.1 A Note on Space Groups Names

Coot uses the space group on the “CRYST1” line of the pdb file. The space group should be one of the xHM symbols listed (for example) in the CCP4 dictionary file syminfo.lib. So, for example, "R 3 2 :H" should be used in preference to "H32".

4.1.2 Read multiple coordinate files

The reading multiple files using the GUI is not available (at the moment). However the following scripting functions are available:

(read-pdb-all)

which reads all the “*.pdb” files in the current directory

(multi-read-pdb glob-pattern dir)

which reads all the files matching glob-pattern in directory dir. Typical usage of this might be:

(multi-read-pdb "a*.pdb" ".")

Alternatively you can specify the files to be opened on the command line when you start coot (see Section Command Line Arguments).

4.1.3 SHELX .ins/.res files

SHELX ".res" (and ".ins" of course) files can be read into Coot, either using the GUI File -> Open Coordinates… or by the scripting function:

(read-shelx-ins-file file-name)

where file-name is quoted, such as "thox.ins".

Although Coot should be able to read any SHELX ".res" file, it may currently have trouble displaying the bonds for centro-symmetric structures.

ShelxL atoms with negative PART numbers are given alternative configuration identifiers in lower case.

To write a SHELX ".ins" file:

(write-shelx-ins-file imol file-name)

where imol is the number of the molecule you wish to export.

This will be a rudimentary file if the coordinates were initially from a "PDB" file, but will contain substantial SHELX commands if the coordinates were initially generated from a SHELX ins file.

4.2 Atom Info

Information about about a particular atom is displayed in the text console when you click using middle-mouse. Information for all the atoms in a residue is available using Info -> Residue Info….

The temperature factors and occupancy of the atoms in a residue can be set by using Edit -> Residue Info….

4.3 Atom Labeling

Use Shift + left-mouse to label atom. Do the same to toggle off the label. The font size is changeable using Edit -> Font Size…. The newly centred atom is labelled by default. To turn this off use:

(set-label-on-recentre-flag 0)

Some people prefer to have atom labels that are shorter, without the slashes and residue name:

(set-brief-atom-labels 1)

To change the atom label colour, use:

(set-font-colour 0.9 0.9 0.9)

4.4 Atom Colouring

The atom colouring system in coot is unsophisticated. Typically, atoms are coloured by element: carbons are yellow, oxygens red, nitrogens blue, hydrogens white and everything else green (see Section Display Manager for colour by chain). However, it is useful to be able to distinguish different molecules by colour, so by default coot rotates the colour map of the atoms (i.e. changes the H value in the HSV ³⁹ colour system). The amount of the rotation depends on the molecule number and a user-settable parameter:

• (set-colour-map-rotation-on-read-pdb 30).

The default value is 31^\circ.

Also one is able to select only the Carbon atoms to change colour in this manner: (set-colour-map-rotation-on-read-pdb-c-only-flag 1).

The colour map rotation can be set individually for each molecule by using the GUI: Edit -> Bond Colours....

4.5 Bond Parameters

The various bond parameters can be set using the GUI dialog Draw -> Bond Parameters or via scripting functions.

The represention style of the molecule that has the active residue (if any) can be changed using the scroll wheel with Ctrl and Shift.

4.5.1 Bond Thickness

The thickness (width) of bonds of individual molecules can be changed. This can be done via the Bond Parameters dialog or the scripting interface:

(set-bond-thickness thickness imol)

where imol is the molecule number.

The default thickness is 3 pixels. The bond thickness also applies to the symmetry atoms of the molecule. The default bond thickness for new molecules can be set using:

(set-default-bond-thickness thick)

where thick is an integer.

There is no means to change the bond thickness of a residue selection within a molecule.

4.5.2 Display Hydrogens

Initially, hydrogens are displayed. They can be undisplayed using

(set-draw-hydrogens mol-no 0) ⁴⁰

where mol-no is the molecule number.

There is a GUI to control this too, under “Edit -> Bond Parameters”.

4.5.3 NCS Ghosts Coordinates

It is occasionally useful when analysing non-crystallographically related molecules to have “images” of the other related molecules appear matched onto the current coordinates. It is important to understand that these ghosts are for displaying differences of NCS-related molecules by structure superposition, not displaying neighbouring NCS related molecules. As you read in coordinates in Coot, they are checked for NCS relationships and clicking on “Edit -> Bond Parameters -> Show NCS Ghosts” -> “Yes” -> “Apply” will create “ghost” copies of them over the reference chain ⁴¹.

Sometimes SSM does not provide a good (or even useful) matrix. In that case, we can specify the residue range ourselves and let the LSQ algorithm provide the matrix. A gui dialog for this operation can be found under Extensions -> NCS -> NCS Ghosts by Residue Range….

The scripting function is used like this:

(manual-ncs-ghosts imol resno-start resno-end ncs-chain-ids)

Typical usage: (manual-ncs-ghosts 0 1 10 (list "A" "B" "C"))

note that in ncs-chain-ids, the NCS master/reference chain-id goes first.

4.5.4 NCS Maps

Coot can use the relative transformations of the NCS-related molecules in a coordinates molecule to transform maps. Use Calculate -> NCS Maps… to do this (note the NCS maps only make sense in the region of the reference chain (see above).

Note also that the internal representation of the map is not transformed. If you try to export a NCS overlay map you will get an untransformed map. A transformed map only makes sense around a given point (and when using transformed maps in Coot, this reference point is changed on the fly, thus allowing map transformations on the fly). [This applies to NCS overlap maps, NCS averaged maps are transformed].

This will also create an NCS averaged map ⁴².

4.5.5 Using Strict NCS

Coot can use a set of strict NCS matrices to specify NCS which means that NCS-related molecules can appear like convention symmetry-related molecules.

(add-strict-ncs-matrix imol ncs-chain-id ncs-target-chain-id m11 m12 m13 m21 m22 m23 m31 m32 m33 t1 t2 t3)

where ncs-chain-id might be "B", "C" "D" (etc.) and ncs-target-chain-id is "A", i.e. the B, C, D molecules are NCS copies of the A chain.

for icosahedral symmetry the translation components t1, t2, t3 will be 0.

You need to turn on symmetry for molecule imol and set the displayed symmetry object type to "Display Near Chains".

4.6 Download coordinates

Coot provides the possibility to download coordinates from an OCA ⁴³. (e.g. EBI) server ⁴⁴ (File -> Get PDB Using Code…). A pop-up entry box is displayed into which you can type a PDB accession code. Coot will then connect to the web server and transfer the file. Coot blocks as it does this (which is not ideal) but on a semi-decent internet connection, it’s not too bad. The downloaded coordinates are saved into a directory called coot-download.

It is also possible to download mmCIF data and generate a map. This currently requires a properly formatted database structure factors mmCIF file ⁴⁵.

4.7 Get Coordinates and Map from EDS

Using this function we have the ability to download coordinates and view the map from structures in the Electron Density Server (EDS) at Uppsala University. This is a much more robust and faster way to see maps from deposited structures. This function can be found under the File menu item.

This feature was added with the assistance of Gerard Kleywegt. If you use the EDS, please cite GJ Kleywegt, MR Harris, JY Zou, TC Taylor, A Wählby & TA Jones (2004), "The Uppsala Electron-Density Server", Acta Cryst. D60, 2240-2249.

4.8 Save Coordinates

On selecting from the menus File -> Save Coordinates… you are first presented with a list of molecules which have coordinates. As well as the molecule number, there is the molecule name - very frequently the name of the file that was read in to generate the coordinates in coot initially. However, this is only a molecule name and should not be confused with the filename to which the coordinates are saved. The coordinates filename can be selected using the Select Filename… button.

If your filename ends in .cif, .mmcif or .mmCIF then an mmCIF file will be written (not a “PDB” file).

4.9 Setting the Space Group

If for some reason, the pdb file that you read does not have a space group, or has the wrong space group, then you can set it using the following function:

(set-space-group imol symbol)

e.g.:

(set-space-group 0 "P 41 21 2")

4.10 Anisotropic Atoms

By default anisotropic atom information is not represented ⁴⁶. To turn them on, use Draw -> Anisotropic Atoms -> Show Anisotropic Atoms? -> Yes, or the command: (set-show-aniso 1).

You cannot currently display thermal ellipsoids ⁴⁷ for isotropic atoms.

4.11 Symmetry

Coordinates symmetry is “dynamic”. Symmetry atoms can be labeled ⁴⁸.

Every time you recentre, the symmetry coordinates are updated. The information shown contains the atom information and the symmetry operation number and translations needed to generate the atom in that position. To show the symmetry operator as a string (rather than a (1-based) index into the list of symmetry operators (as is the default)) Use Draw -> Show Symmetry> -> Expanded Symmetry Atom Labels. Coot generates symmetry related atoms by moving the current set close to the origin by a translation, performing the symmetry expansion around the origin and moving the the symmetry coordinates back by applying the inverse of the origin translation. The origin translation is also displayed in curly braces, e.g. “{1 -1 0}”.

By default symmetry atoms are not displayed.

If you want coot to display symmetry coordinates without having to use the gui, add to your ~/.coot the following:

(set-show-symmetry-master 1)

The symmetry can be represented as C\alphas. This along with representation of the molecule as C\alphas (Section Display Manager) allow the production of a packing diagram.

4.11.1 Missing symmetry

Sometimes (rarely) coot misses symmetry-related molecules that should be displayed. In that case you need to expand the shift search (the default is 1):

(set-symmetry-shift-search-size 2)

This is a hack, until the symmetry search algorithm is improved.

4.12 Sequence View

The protein is represented by one letter codes and coloured according to secondary structure. These one letter codes are active - if you click on them, they will change the centre of the graphics window - in much the same way as clicking on a residue in the Ramachandran plot.

4.13 Print Sequence

The single letter code (of the imolth molecule) is written out to the console in FASTA format. Use can use this to cut and paste into other applications:

(print-sequence imol)

4.14 Environment Distances

Environment distances are turned on using Info -> Environment Distances…. Contacts to other residues are shown and to symmetry-related atoms if symmetry is being displayed. The contacts are coloured by atom type ⁴⁹.

4.15 Distances and Angles

The distance between atoms can be found using Info -> Distance ⁵⁰. The result is displayed graphically, and written to the console.

4.16 Zero Occupancy Marker

Atoms of zero occupancy are marked with a grey spot. To turn off these markers, use:

(set-draw-zero-occ-markers 0)

Use an argument of 1 to turn them on.

4.17 Atomic Dots

You can draw dots round arbitrary atom selections

(dots imol atom-selection dot-density radius) The function returns a handle.

e.g. put a sphere of dots around all atoms of the 0th molecule (it might be a set of heavy atom coordinates) at the default dot density and radius:

(dots 0 "/1" "heavy-atom-sites" 1 1)

You can’t change the colour of the dots.

There is no internal mechanism to change the radius according to atom type. With some cleverness you might be able to call this function several times and change the radius according to the atom selection.

There is a function to clear up the dots for a particular molecule imol and dots set identifier dots-handle

(clear-dots imol dots-handle)

There is a function to return how many dots sets there are for a particular molecule imol:

(n-dots-set imol)

4.18 Ball and Stick Representation

Fragments of the molecule can be rendered as a “ball and stick” molecule:

(make-ball-and-stick imol atom-selection bond-thickness sphere-size draw-spheres-flag)

e.g. (make-ball-and-stick 0 "/1/A/10-20" 0.3 0.4 1)

The ball-and-stick representation can be cleared using:

(clear-ball-and-stick imol)

4.19 Mean, Median Temperature Factors

Coot can be used to calculate the mean (average) and median temperatures factors:

(average-temperature-factor imol)

(median-temperature-factor imol)

-1 is returned if there was a problem ⁵¹.

4.20 Secondary Structure Matching (SSM)

The excellent SSM alogrithm⁵² of Eugene Krissinel is available in Coot. The GUI interface is straight-forward and can be found under Calculate -> SSM Superpose. You can specify the specific chains that you wish to match using the "Use Specific Chain" check-button.

There is a scripting level function which gives even finer control:

(superpose-with-atom-selection imol1 imol2 mmdb-atom-selection-string-1 mmdb-atom-selection-string-2 move-copy-flag )

the move-copy-flag should be 1 if you want to apply the transformation to a copy of imol2 (rather than imol2 itself). Otherwise, move-copy-flag should be 0.

mmdb atom selection strings (Coordinate-IDs) are explained in detail in the mmdb manual.

Briefly, the string should be formed in this manner:

/mdl/chn/seq(res).ic/atm[elm]:aloc

e.g. "/1/A/12-130/CA"

<p><a href="http://www.ebi.ac.uk/~keb/cldoc/object/cl_obj_surf.html#CoordinateID">The mmdb manual CoordinateID description</a>.</p>

4.21 Least-Squares Fitting

There is a simple GUI for this Calculate -> LSQ Superpose…

The scripting interface to LSQ fitting is as follows:

(simple-lsq-match ref-start-resno ref-end-resno ref-chain-id imol-ref mov-start-resno mov-end-resno mov-chain-id imol-mov match-type)

where:

• ref-start-resno is the starting residue number of the reference molecule
• ref-end-resno is the last residue number of the reference molecule
• mov-start-resno is the starting residue number of the moving molecule
• mov-end-resno is the last residue number of the moving molecule
• match-type is one of 'CA, 'main, or 'all.

e.g.: (simple-lsq-match 940 950 "A" 0 940 950 "A" 1 'main)

More sophisticated (match molecule number 1 chain “B” on to molecule number 0 chain “A”):

(define match1 (list 840 850 "A" 440 450 "B" 'all))

(define match2 (list 940 950 "A" 540 550 "B" 'main))

(clear-lsq-matches)

(set-match-element match1)

(set-match-element match2)

(apply-lsq-matches 0 1) ; match molecule number 1 onto molecule number 0.

4.22 Ligand Overlaying

The scripting function

(overlap-ligands imol-ligand imol-ref chain-id-ref resno-ref)

returns a rotation+translation operator which can be applied to other molecules (and maps). Here, imol-ligand is the molecule number of the ligand (which is presumed to be a a molecule on its own - Coot simply takes the first residue that it finds). imol-ref chain-id-ref resno-ref collectively describe the target position for the moving imol-ligand molecule.

The convenience function

(overlay-my-ligands imol-mov chain-id-mov resno-mov imol-ref chain-id-ref resno-ref)

wraps overlap-ligands.

The GUI for the function can be found under

Extensions -> Modelling -> Supperpose Ligands…

4.23 Writing PDB files

As well as the GUI option File -> Save Coordinates… there is a scripting options available:

(write-pdb-file imol pdb-file-name)

which writes the imolth coordinates molecule to filename.

To write a specific residue range:

(write-residue-range-to-pdb-file imol chain-id start-resno endresno pdb-file-name)

5 Modelling and Building

The functions described in this chapter manipulate, extend or build molecules and can be found under Calculate -> Model/Fit/Refine…. When activated, the dialog "stays on top" of the main graphics window ⁵³. Some people think that this is not always desirable, so this behaviour can be undone using:

(set-model-fit-refine-dialog-stays-on-top 0)

5.1 Regularization and Real Space Refinement

Coot will read the geometry restraints for refmac and use them in fragment (zone) idealization - this is called “Regularization”. The geometrical restraints are, by default, bonds, angles, planes and non-bonded contacts. You can additionally use torsion restraints by Calculate -> Model/Fit/Refine… -> Refine/Regularize Control -> Use Torsion Restraints. Truth to tell, this has not been successful in my hands (sadly).

“RS (Real Space) Refinement” (after Diamond, 1971 ⁵⁴) in Coot is the use of the map in addition to geometry terms to improve the positions of the atoms. Select “Regularize” from the “Model/Fit/Refine” dialog and click on 2 atoms to define the zone (you can of course click on the same atom twice if you only want to regularize one residue). Coot then regularizes the residue range. At the end Coot, displays the intermediate atoms in white and also displays a dialog, in which you can accept or reject this regularization. In the console are displayed the \chi^2 values of the various geometrical restraints for the zone before and after the regularization. Usually the \chi^2 values are considerably decreased - structure idealization such as this should drive the \chi^2 values toward zero.

The use of “Refinement” is similar - with the addition of using a map. The map used to refine the structure is set by using the “Refine/Regularize Control” dialog. If you have read/created only one map into Coot, then that map will be used (there is no need to set it explicitly).

Use, for example, (set-matrix 20.0)

to change the weight of the map gradients to geometric gradients. The higher the number the more weight that is given to the map terms ⁵⁵. The default is 60.0. This will be needed for maps generated from data not on (or close to) the absolute scale or maps that have been scaled (for example so that the sigma level has been scaled to 1.0).

For both “Regularize Zone” and “Refine Zone” one is able to use a single click to refine a residue range. Pressing A on the keyboard while selecting an atom in a residue will automatically create a residue range with that residue in the middle. By default the zone is extended one residue either size of the central residue. This can be changed to 2 either side using (set-refine-auto-range-step 2).

Intermediate (white) atoms can be moved around with the mouse (click and drag with left-mouse, by default). Refinement will proceed from the new atom positions when the mouse button is released. It is possible to create incorrect atom nomenclature and/or chiral volumes in this manner - so some care must be taken. Press the A key as you left-mouse click to move atoms more “locally” (rather than a linear shear) and Ctrl key as you left-mouse click to move just one atom.

In more up to date versions, Coot will display colour patches (something like a traffic light system) representing the chi squared values of each of types of geometric feature refined. Typically “5 greens” is the thing to aim for, the colour changes occurring at chi squared values 2, 5 and 8 (8 being the most red).

To prevent the unintentional refinement of a large number of residues, there is a “heuristic fencepost” of 20 residues. A selection of than 20 residues will not be regularized or refined. The limit can be changed using the scripting function: e.g. (set-refine-max-residues 30).

5.1.1 Dictionary

The geometry description for residues, monomers and links used by Coot are in the standard mmCIF format. Because this format alows multiple comp_ids (residue types) to be described within a cif loop, it is hard to tell when a dictionary entry needs to be overwritten when reading a new file. Therefore Coot makes this extra constraint: that the “chem_comp” loop should appear first in the comp list data item - if this is the case, then Coot can overwrite an old restraint table for a particular comp_id/residue-type when a new one is read.

By default, the geometry dictionary entries for only the standard residues are read in at the start ⁵⁶. It may be that your particular ligand is not amongst these. To interactively add a dictionary entry use File -> Import CIF Dictionary.

There is a selector in the cif dictionary file chooser that allow you to select the molecule to which molecule refers. Each of the indididual molecules can be specifically selected. “All“ means that the dictionary refers to all loaded molecules and “Auto“ means that the dictionary will be applied to all model molecules if the comp_id/residue name is not on the non-auto-load list and will choose the latest model molecule if the comp_id/residue name is on the non-auto-load list. By default the non-auto-load list consists of INH, LIG, DRG, XXX, and the series LG0-9.

Alternatively, you can use the function:

(read-cif-dictionary filename)

and add this to your .coot file (this may be the preferred method if you want to read the file on more than one occasion).

Note: the dictionary also provides the description of the ligand’s torsions.

5.1.2 Sphere Refinement

Sphere refinement selects residues within a certain distance of the residue at the centre of the screen and includes them for real space refinement. In this way, one can select residues that are not in a linear range. This technique is useful for refining disulfide bonds and glycosidic linkages.

To enable sphere refinement, Right-mouse click in the right hand side of the horizontal toolbutton menu, Manage buttons -> [Tick] Sphere Refine -> Apply. You will need a python-enabled Coot to do this.

The following adds a key binding (Shift-R) that refines resides that are within 3.5Å of the residue at the centre of the screen:

        
(define *sphere-refine-radius* 3.5)

(add-key-binding "Refine residues in a sphere" "R"
    (lambda ()
     (using-active-atom
      
      (let* ((rc-spec (list aa-chain-id aa-res-no aa-ins-code))
	     (ls (residues-near-residue aa-imol rc-spec *sphere-refine-radius*)))
	(refine-residues aa-imol (cons rc-spec ls))))))

5.1.3 Refining Specific Residues

You can specify the residues that you want to refine without using a linear or sphere selection usine refine-residues. For example:

(refine-residues 0 '(("L" 501 "") ("L" 503 "")))

will refine residues A501 and A503 (and residue A502 (if it exists) will be an anchoring residue - used in optimizing the link geometry of the atoms in A501 and A503).

5.1.4 Refining Carbohydrates

Refining carbohydrates monomers should be as straightforward as refining a protein residue. Coot will look in the dictionary for the 3-letter code for the particular residue type, if it does not find it, Coot will try to search for dictionary files using “-b-D” or “-a-L” extensions.

When refining a group of carbohydrates, the situation needs a bit more explanation. For each residue pair with tandem residue numbers specified in the refinement range selection, Coot checks if these residue types are are furanose or pyranose in the dictionary, and if the are both one or the other, then it tries to see if there are any of the 11 link types (BETA1-4, BETA2-3, ALPHA1-2 and so on) specified in the dictionary. It does this by a distance check of the potentially bonding atoms. If the distance is less than 3.0Å, then a glycosidic bond is made and used in the refinement.

Bonds between protein and carbohydrate and branched carbohydrates can be refined using “Sphere Refinement”.

Instead of using a sphere to make a residue selection, you can specify the residues directly using refine-residues, for example:

(refine-residues 0 '(("L" 501 "") ("L" 503 "")))

LINK and LINKR cards are not yet used to determine the geometry of the restraints.

5.1.5 Planar Peptide Restraints

By default, Coot uses a 5 atom (CA-1, C-1, O-1, N-2, CA-2) planar peptide restraints. These restraints should help in low resolution fitting (the main-chains becomes less distorted), reduce accidental cis-peptides and may help “clean up” Ramachandran plots.

(add-planar-peptide-restraints)

And similarly they can be removed:

(remove-planar-peptide-restraints)

There is also a GUI to add and remove these restraints in Extensions -> Refine… -> Peptide Restraints…

5.1.6 The UNK residue type

The UNK residue type is a special residue type to Coot. It has been added for use with Buccaneer. Don’t give you ligand (or anything else) the 3-letter-code UNK or confusion will result ⁵⁷.

5.1.7 Moving Zero Occupancy Atoms

By default, atoms with zero occupancy are moved when refining and regularizing. This can sometimes be inconvenient. To turn of the movement of atoms with zero occupancy when refining and regularizing:

(set-refinement-move-atoms-with-zero-occupancy 0)

5.2 Changing the Map for Building/Refinement

You can change the map that is used for the fitting and refinement tools using the Select Map... button on the Model/Fit/Refine dialog.

5.3 Rotate/Translate Zone

“Rotate/Translate Zone” from the “Model/Fit/Refine” menu allows manual movement of a zone. After pressing the “Rotate/Translate Zone” button, select two atoms in the graphics canvas to define a residue range ⁵⁸, the second atom that you click will be the local rotation centre for the zone. The atoms selected in the moving fragment have the same alternate conformation code as the first atom you click. To actuate a transformation, click and drag horizontally across the relevant button in the newly-created “Rotation & Translation” dialog. The axis system of the rotations and translations are the screen coordinates. Alternatively ⁵⁹, you can click using left-mouse on an atom in the fragment and drag the fragment around. Use Control Left-mouse to move just one atom, rather than the whole fragment. If you click Control Left-mouse whilst not over an atom then you can rotate the fragment using mouse drag. Click “OK” (or press Return) when the transformation is complete.

To change the rotation point to the centre of the intermediate atoms (rather than the second clicked atom), use the setting:

(set-rotate-translate-zone-rotates-about-zone-centre 1)

5.4 Rigid Body Refinement

“Rigid Body Fit Zone” from the “Model/Fit/Refine” dialog provides rigid body refinement. The selection is zone-based ⁶⁰. So to refine just one residue, click on one atom twice.

Sometimes no results are displayed after Rigid Body Fit Zone. This is because the final model positions had too many final atom positions in negative density. If you want to over-rule the default fraction of atoms in the zone that have an acceptable fit (0.75), to be (say) 0.25:

(set-rigid-body-fit-acceptable-fit-fraction 0.25)

5.5 Simplex Refinement

Rigid body refinement via Nelder-Mead Simplex minimization is available in Coot. Simplex refinement has a larger radius of convergence and thus is useful in a position where simple rigid body refinement finds the wrong minimum. However the Simplex algorithm is much slower. Simplex refinement for a residue range start-resno to end-resno (inclusive) in chain chain-id can be accessed as follows:

(fit-residue-range-to-map-by-simplex start-resno end-resno alt-loc chain-id imol imol-for-map)

There is currently no GUI interface to Simplex refinement.

5.6 Post-manipulation-hook

If you wanted automatically run a function after a model has been manipulated then you can do so using by creating a function that takes 2 arguments, such as:

(post-manipulation-hook imol manipulation-mode)

manipulation-mode is one of (DELETED), (MUTATED) or (MOVINGATOMS).

And of course imol is the model number of the maniplated molecule.

(It would of course be far more useful if this function was also passed a list of residues - that is something for the future).

5.7 Baton Building

Baton build is most useful if a skeleton is already calculated and displayed (see Section Skeletonization). When three or more atoms have been built in a chain, Coot will use a prior probability distribution for the next position based on the position of the previous three. The analysis is similar to that of Oldfield & Hubbard (1994) ⁶¹, however it is based on a more recent and considerably larger database.

Little crosses are drawn representing directions in which is is possible that the chain goes, and a baton is drawn from the current point to one of these new positions. If you don’t like this particular direction ⁶², use Try Another. The list of directions is scored according to the above criterion and sorted so that the most likely is at the top of the list and displayed first as the baton direction.

When starting baton building, be sure to be about 3.8Å from the position of the first-placed C\alpha, this is because the next C\alpha is placed at the end of the baton, the baton root being at the centre of the screen. So, when trying to baton-build a chain starting at residue 1, centre the screen at about the position of residue 2.

It seems like a good idea to increase the map sampling to 2 or even 2.5 (before reading in your mtz file) [a grid sampling of about 0.5Å seems reasonable] when trying to baton-build a low resolution map. You can set the map sampling using Edit -> Map Parameters -> Map Sampling.

Occasionally, every point is not where you want to position the next atom. In that case you can either shorten or lengthen the baton, or position it yourself using the mouse. Use “b” on the keyboard to swap to baton mode for the mouse ⁶³.

Baton-built atoms are placed into a molecule called “Baton Atom” and it is often sensible to save the coordinates of this molecule before quitting coot.

If you try to trace a high resolution map (1.5Å or better) you will need to increase the skeleton search depth from the default (10), for example:

(set-max-skeleton-search-depth 20)

Alternatively, you could generate a new map using data to a more moderate resolution (2Å), the map may be easier to interpret at that resolution anyhow ⁶⁴.

The guide positions are updated every time the “Accept” button is clicked. The molecule name for these atoms is “Baton Build Guide Points” and is is not usually necessary to keep them.

5.7.1 Undo

There is also an “Undo” button for baton-building. Pressing this will delete the most recently placed C\alpha and the guide points will be recalculated for the previous position. The number of “Undo”s is unlimited. Note that you should use the “Undo” button in the Baton Build dialog, not the one in the “Model/Fit/Refine” dialog (Section Backups and Undo).

5.7.2 Missing Skeleton

Sometimes (especially at loops) you can see the direction in which the chain should go, but there is no skeleton (see Section Skeletonization) is displayed (and consequently no guide points) in that direction. In that case, “Undo” the previous atom and decrease the skeletonization level (Edit -> Skeleton Parameters -> Skeletonization Level). Accept the atom (in the same place as last time) and now when the new guide points are displayed, there should be an option to build in a new direction.

5.7.3 Building Backwards

The following scenario is not uncommon: you find a nice stretch of density and start baton building in it. After a while you come to a point where you stop (dismissing the baton build dialog). You want to go back to where you started and build the other way. How do you do that?

• Use the command:

  (set-baton-build-params start-resno chain-id "backwards")

  where start-resno would typically be 0 ⁶⁵ and chain-id would be "" (default).

• Recentre the graphics window on the first atom of the just-build fragment
• Select “Ca Baton Mode” and select a baton direction that goes in the “opposite” direction to what is typically residue 2. This is slightly awkward because the initial baton atoms build in the “opposite” direction are not dependent on the first few atoms of the previously build fragment.

5.8 Reversing Direction of Fragment

After you’ve build a fragment, sometimes you might want to change the direction of that fragment (this function changes an already existing fragment, as opposed to Backwards Building which sets up Baton Building to place new points in reverse order).

The fragment is defined as a contiguous set of residues numbers. So that you should be sure that other partial fragments which have the same chain id and that are not connected to this fragment have residue numbers that are not contiguous with the fragment you are trying to reverse.

5.9 C\alpha -> Mainchain

Mainchain can be generated using a set of C\alphas as guide-points (such as those from Baton-building) along the line of Esnouf ⁶⁶ or Jones and coworkers ⁶⁷. Briefly, 6-residue fragments of are generated from a list of high-quality ⁶⁸ structures. The C\alpha atoms of these fragments are matched against overlapping sets of the guide-point C\alphas. The resulting matches are merged to provide positions for the mainchain (and C\beta) atoms. This procedure works well for helices and strands, but less well ⁶⁹ for less common structural features.

This function is also available from the scripting interface:

(db-mainchain imol chain-id resno-start resno-end direction)

where direction is either "backwards" or "forwards".

Recall that the chain-id needs to be quoted, i.e. use "A" not A. Note that chain-id is "" when the C\alphas have been built with Baton Mode in Coot.

5.10 Backbone Torsion Angles

It is possible to edit the backbone \phi and \psi angles indirectly using an option in the Model/Fit/Refine’s dialog: “Edit Backbone Torsions..”. When clicked and an atom of a peptide is selected, this produces a new dialog that offers “Rotate Peptide” which changes this residues \psi and “Rotate Carbonyl” which changes \phi. Click and drag across the button ⁷⁰ to rotate the moving atoms in the graphics window. You should know, of course, that making these modifications alter the \phi/\psi angles of more than one residue.

5.11 Docking Sidechains

Docking sidechains means adding sidechains to a model or fragment that has currently only poly-Ala, where the sequence assignment is unknown. The algorithm is basically the same as in Cowtan’s Buccaneer, but with some corners cut to make things (more or less) interactive. The algorithm uses the shape of the density around the C-beta position to estimate the probability of each sidechain type at that position.

The function is accessed via the Extensions -> Dock Sequence menu item. First, a sequence should be assigned from a PIR file to a particular chain-id and model number. Secondly Extensions -> Dock Sequence -> Dock Sequence on this fragment…. Choose the model to build on and then Dock Sequence! If all goes well, the model will be updated with mutated residues and undergo rotamer seach for each of the new residues. If the sequence alignment is not sufficiently clear, then you will get a dialog suggesting that you extend or improve the fragment.

5.12 Rotamers

The rotamers are generated ⁷¹ from the backbone independent sidechain library of the Richardsons group ⁷².

The m, t and p stand for “minus (-60)”, “trans (180)” and “plus (+60)”. There is one letter per \chi angle.

Use keyboard . and , to cycle round the rotamers.

5.12.1 Auto Fit Rotamer

“Auto Fit Rotamer” will try to fit the rotamer to the electron density. Each rotamer is generated, rigid body refined and scored according to the fit to the map. Fitting the second conformation of a dual conformation in this way will often fail - the algorithm will pick the best fit to the density - ignoring the position of the other atoms.

The algorithm doesn’t know if the other atoms in the structure are in sensible positions. If they are, then it is sensible not to put this residue too close to them, if they are not then there should be no restriction from the other atoms as to the position of this residue - the default is “are sensible”, which means that the algorithm is prevented from finding solutions that are too close to the atoms of other residues. (set-rotamer-check-clashes 0) will stop this.

There is a scripting interface to auto-fitting rotamers:

(auto-fit-best-rotamer resno alt-loc ins-code chain-id imol-coords imol-map clash-flag lowest-rotamer-probability)

where:

resno is the residue number

alt-loc is the alternate/alternative location symbol (e.g. "A" or "B", but most often "")

ins-code is the insertion code (usually "")

imol-coords is the molecule number of the coordinates molecule

imol-map is the molecule number of the map to which you wish to fit the side chains

clash-flag should the positions of other residues be included in the scoring of the rotamers (i.e. clashing with other other atoms gets marked as bad/unlikely)

lowest-rotamer-probability: some rotamers of some side chains are so unlikely that they shouldn’t be considered - typically 0.01 (1%).

You can change the auto-fit rotamer fitting algorithms using

(set-rotamer-search-mode mode)

where mode is one of (ROTAMERSEARCHAUTOMATIC), (ROTAMERSEARCHLOWRES) (i.e. "Backrub Rotamers" (vide infra)) or (ROTAMERSEARCHHIGHRES) (the conventional/high-resolution method using rigid-body fitting).

By default, the auto-fit rotamer method is (ROTAMERSEARCHAUTOMATIC).

5.12.1.1 Backrub Rotamers

By default, Auto Fit Rotamer will switch to “Backrub Rotamer” ⁷³ mode when fitting against a map of worse than 2.7Å. This search mode moves the some atoms of the mainchain of the neighbouring residues. After rotation of the central residue and neighbouring atoms around the “backrub vector”, the individual peptides are back-rotated (along the peptide axis) so that the carbonyl oxygen are placed as near as possible to their original position. The Ramachandran plot is not used in this fitting algorithm.

5.12.2 De-clashing residues

Sometimes you don’t have a map ⁷⁴ but nevertheless there are clashing residues ⁷⁵ (for example after mutation of a residue range) and you need to rotate side-chains to a non-clashing rotamer. There is a scripting interface:

(de-clash imol chain-id start-resno end-resno)

start-resno is the residue number of the first residue you wish to de-clash

end-resno is the residue number of the last residue you wish to de-clash

imol is the molecule number of the coordinates molecule

This interface will not change residues with insertion codes or alternate conformation. The lowest-rotamer-probability is set to 0.01.

5.13 Editing chi Angles

Instead of using Rotamers, one can instead change the \chi angles (often called “torsions”) “by hand” (using “Edit Chi Angles” from the “Model/Fit/Refine” dialog). To edit a residue’s \chi_1 press “1”: to edit \chi_2, “2”: \chi_3 “3” and \chi_4 “4”. Use left-mouse click and drag to change the \chi value. Use keyboard “0” ⁷⁶ to go back to ordinary view mode at any time during the editing. Alternatively, one can use the “View Rotation Mode” or use the Ctrl key when moving the mouse in the graphics window. Use the Accept/Reject dialog when you have finished editing the \chi angles.

For non-standard residues, the clicked atom defines the base of the atom tree, which defines the “head” of the molecule (it’s the “tail” (twigs/leaves) that wags). To emphasise, then: it matters on which atom you click!

By default torsions for hydrogen atoms are turned off. To turn them on:

(set-find-hydrogen-torsions 1)

To edit the rotatable bonds of a ligand using this tool, you will need to have read in the mmCIF dictionary beforehand.

5.14 Torsion General

You need to click on the torsion-general button, then click 4 atoms that describe the torsion - the first atom will be the base (non moving) part of the atom tree, on clicking the 4th atom a dialog will pop up with a "Reverse" button. Move this dialog out of the way and then left mouse click and drag in the main window will rotate the "top" part of the residue round the clicked atoms 2 and 3. When you are happy, click "Accept".

If you are torsion generaling a residue that has an alt conf, then the atoms of residue that are moved are those that have the same alt conf as the 4th clicked atom (or have an blank alt conf).

5.14.1 Ligand Torsion angles

For ligands, you will need to read the mmCIF file that contains a description of the ligand’s geometry (see Section Regularization and Real Space Refinement). By default, torsions that move hydrogens are not included. Only 9 torsion angles are available from the keyboard torsion angle selection.

5.15 Pep-flip

Coot uses the same pepflip scheme as is used in O (i.e. the C, N and O atoms are rotated 180^o round a line joining the C\alpha atoms of the residues involved in the peptide). Flip the peptide again to return the atoms to their previous position.

5.16 Add Alternate Conformation

The allows the addition alternate (dual, triple etc.) conformations to the picked residue. By default, this provides a choice of rotamer (Section Rotamers). If there are not the correct main chain atoms a rotamer choice cannot be provided, and Coot falls back to providing intermediate atoms.

The default occupancy for new atoms is 0.5. This can be changed by using use slider on the rotamer selection window or by using the scripting function:

(set-add-alt-conf-new-atoms-occupancy 0.4)

The remaining occupancy of the atoms (after the new occupancy has been added) is split amongst the atoms that existed in the residue before the split. It is important therefore that the residues atoms have sane occupancies before adding an alternative conformation.

The default Split Type is to split the whole residue. If you want the default to be to split a residue after (and including) the CA, then add to your .coot file:

(set-add-alt-conf-split-type-number 0)

5.17 Mutation

Mutations are available on a 1-by-1 basis using the graphics. After selecting “Mutate…” from the “Model/Fit/Refine” dialog, click on an atom in the graphics. A “Residue Type” window will now appear. Select the new residue type you wish and the residue in the graphics is updated to the new residue type ⁷⁷. The initial position of the new rotamer is the a priori most likely rotamer. Note that in interactive mode, such as this, a residue type match ⁷⁸ will not stop the mutation action occurring.

5.17.1 Mutating DNA/RNA

Mutation of DNA or RNA can be performed using “Simple Mutate” from the Model/Fit/Refine dialog. Residues need to be named "Ad", "Gr", "Ur" etc.

5.17.2 Multiple mutations

This dialog can be found under Calculate -> Mutate Residue Range. A residue range can be assigned a sequence and optionally fitted to the map. This is useful converting a poly-ALA model to the correct sequence ⁷⁹.

Multiple mutations are also supported via the scripting interface. Unlike the single residue mutation function, a residue type match will prevent a modification of the residue ⁸⁰. Two functions are provided: To mutate a whole chain, use (mutate-chain imol chain-id sequence) where:

chain-id is the chain identifier of the chain that you wish to mutate (e.g. "A") and

imol is molecule number.

sequence is a list of single-letter residue codes, such as "GYRESDF" (this should be a straight string with no additional spaces or carriage returns).

Note that the number of residues in the sequence chain and those in the chain of the protein must match exactly (i.e. the whole of the chain is mutated (except residues that have a matching residue type).)

To mutate a residue range, use

• (mutate-residue-range imol chain-id start-res-no stop-res-no sequence)

where

start-res-no is the starting residue for mutation

stop-res-no is the last residue for mutation, i.e. using values of 2 and 3 for start-res-no and stop-res-no respectively will mutate 2 residues.

Again, the length of the sequence must correspond to the residue range length. Note also that this is a protein sequence - not nucleic acid.

For mutation of nucleic acids, use:

(mutate-nucleotide-range imol chain-id resno-start resno-end sequence)

5.17.3 Mutating to a Non-Standard Residue

Sometimes one might like to model post-translational or other such modifications. How is that done, if the new residue type is not one of the standard residue types?

There is a scripting function:

(mutate-by-overlap imol chain-id resno new-three-letter-code)

This imports a model residue for the new residue type and overlays it on to the given residue by using graph-matching to determine the equivalent atoms.

The GUI for this can be found under Extensions -> Modelling -> Replace Residue... (for this to work, you need to be centred on the residue you wish to replace).

Note that if you are replacing are conventional protein residue with a modified form (e.g. replacing a TYR with a phoso-tyrosine or a LYS with an acetyl-lysine) you will need to make sure that the group of the resulting restraints is an L-peptide (use Edit -> Restraints to check and modify the restraints group. Likewise for modified RNA/DNA nucleotides, you need to specify the group as RNA or DNA as appropriate.

5.17.4 Mutate and Autofit

The function combines Mutation and Auto Fit Rotamer and is the easiest way to make a mutation and then fit to the map. You can currently only “Mutate and Autofit” protein residues (i.e. things with a rotamer dictionary.

5.17.5 Renumbering

Renumbering is straightforward using the renumber dialog available under Calculate -> Renumber Residue Range…. There is also a scripting interface:

(renumber-residue-range imol chain-id start-res-no last-resno offset)

5.18 Importing Lignds/Monomers

You can import monomers (often ligands) using File -> Get Monomer…⁸¹ by providing the 3-letter code of your monomer/ligand. The resulting molecule will be moved so that it placed at the current screen centre.

Typically, when you are happy about the placement of the ligand, you’d then use Merge Molecules to add the ligand/monomer to the main set of coordinates.

This procedure creates a pdb file monomer-XXX.pdb and a dictionary file libcheck_XXX.cif in the directory in which Coot was started.

A future invocation of Get Monomer uses these file so that the monomer appears quickly ⁸².

5.19 Ligand from SMILES strings

Similarly, you can generate ligands using File -> SMILES... and providing a SMILES string and a code for the residue name (this is your name for the residue type and a dictionary will be generated for the monomer of this type). This function is also a wrapper to LIBCHECK.

5.20 Find Ligands

You are offered a selection of maps to search (you can only choose one at a time) and a selection of molecules that act as a mask to this map. Finally you must choose which ligand types you are going to search for in this map ⁸³. Only molecules with less than 400 atoms are suggested as potential ligands.

If you do not have any molecules with less that 400 atoms loaded in Coot, you will get the message:

"Error: you must have at least one ligand to search for!"

New ligands are placed where the map density is and protein (mask) atoms are not). The masked map is searched for clusters using a default cut-off of 1.0\sigma. In weak density this cut-off may be too high and in such a case the cut-off value can be changed using something such as:

(set-ligand-cluster-sigma-level 0.8)

However, if the map to be searched for ligands is a difference map, a cluster level of 2.0 or 3.0 would probably be more appropriate (less likely to generate spurious sites).

Each ligand is fitted with rigid body refinement to each potential ligand site in the map and the best one for each site selected and written out as a pdb file. The clusters are sorted by size, the biggest one first (with an index of 0). The output placed ligands files have a prefix “best-overall” and are tagged by the cluster index and residue type of the best fit ligand in that site.

By default, the top 10 sites are tested for ligands - to increase this use:

(set-ligand-n-top-ligands 20)

5.20.1 Flexible Ligands

If the “Flexible?” checkbutton is activated, coot will generate a number of variable conformations (default 100) by rotating around the rotatable bonds (torsions). Each of these conformations will be fitted to each of the potential ligand sites in the map and the best one will be selected (again, if it passes the fitting criteria above).

Before you search for flexible ligands you must have read the mmCIF dictionary for that particular ligand residue type (File -> Import CIF dictionary).

Use:

(set-ligand-flexible-ligand-n-samples n-samples)

where n-samples is the number of samples of flexibility made for each ligand. Generally speaking, The more the number of rotatable bonds, the bigger this number should be.

By default the options to change these values are not in the GUI. To enable these GUI options, use the scripting function:

(ligand-expert)

5.20.2 Adding Ligands to Model

After successful ligand searching, one may well want to add that displayed ligand to the current model (the coordinates set that provided the map mask). To do so, use Merge Molecules (Section Merge Molecules).

5.21 Flip Ligand

Sometimes a ligand is placed more or less in the correct position, but the orientation is wrong - or at least you might want to explore other possible orientation. To do that easily a function has been provided:

(flip-ligand imol chain-id residue-number)

This will flip the orientation of the residue around the Eigen vector corresponding to the largest Eigen value, exploring 4 possible orientations.

This function has been further wrapped to provide flipping for the active residue:

(flip-active-ligand)

This function can easily be bound to a key.

5.22 Find Waters

As with finding ligands, you are given a choice of maps, protein (masking) atoms. A final selection has to be made for the cut-off level, note that this value is the number of standard deviation of the density of the map before the map has been masked. The default sigma level (water positions must have density above this level) is set for a “2Fo-Fc”-style map. If you want to use a difference map, you must change the sigma level (typically to 3 sigma) otherwise you run the risk of fitting waters to difference map noise peaks.

Then the map is masked by the masking atoms and a search is made of features in the map about the electron density cut-off value. Waters are added if the feature is approximately water-sized and can make sensible hydrogen bonds to the protein atoms. The new waters are optionally created in a new molecule called “Waters”.

You have control over several parameters used in the water finding:

(set-write-peaksearched-waters)

which writes ligand-waters-peaksearch-results.pdb, which contains the water peaks (from the clusters) without any filtering and ligand-waters.pdb which are a disk copy filtered waters that have been either added to the molecule or from which a new molecule has been created.

(set-ligand-water-to-protein-distance-limits min-d max-d) sets the minimum and maximum allowable distances between new waters and the masking molecule (usually the protein). Defaults are 2.4 and 3.2Å.

(set-ligand-water-spherical-variance-limit varlim) sets the upper limit for the density variance around water atoms. The default is 0.12.

The map that is marked by the protein and is searched to find the waters is written out in CCP4 format as "masked-for-waters.map".

5.22.1 Refinement Failure

Sometimes as a result of water fitting, you may see something like:

 WARNING:: refinement failure
           start pos: xyz = (      17.1,     34.76,     60.42)
           final pos: xyz = (     17.19,     34.61,     60.59)

When Coot finds a blob, it does a crude positioning of an atom at the centre of the grid points. It then proceeds to move to the peak of the blob by a series of translations. There are a certain number of cycles, and if it doesn’t reach convergence by the end of those cycles then you get the error message.

Often when you go to the position indicated, you can see why Coot had a problem in the refinement.

5.22.2 Blobs

After a water search, Coot will create a blobs dialog (see Section sec_blobs).

5.23 Add Terminal Residue

This creates a new residue at the C or N terminal extension of the residue clicked by fitting to the map. \phi,\psi angle pairs are selected at random based on the Ramachandran plot probability (for a generic residue) and fitted to the density. By default there are 100 trials. It is possible that a wrong position will be selected for the terminal residue and if so, you can reject this fit and try again with Fit Terminal Residue ⁸⁴. Each of the trial positions are scored according to their fit to the map ⁸⁵ and the best one selected. It is probably a good idea to run “Refine Zone” on these new residues.

If you use the Extensions (Dock Sequence... -> Associate Sequence with Model) to apply a PIR sequence file to a model then Add Terminal Residue will use the sequence alignment to determine the residue type of the added residue.

Sometimes, particularly with low resolution maps, the added terminal residue will wander off to somewhere inappropriate. This can be addressed in a number of ways:

1. (set-terminal-residue-do-rigid-body-refine 0) will disable rigid body fitting of the terminal residue fragment for each trial residue position (the default is 1 (on)) - this may help if the search does not provide good results.
2. to anneal the newly added residue back to the clicked residue (no matter where it ended up being positioned): (set-add-terminal-residue-do-post-refine 1)
3. (set-add-terminal-residue-n-phi-psi-trials 200) will change the number of trials (default is 100). This is useful if you think that Coot needs to search harder to find a good solution to the positioning of the next residue.

5.24 Add OXT Atom to Residue

At the C-terminus of a chain of amino-acid residues, there is a “modification” so that the C-O becomes a carbonyl, i.e. an extra (terminal) oxygen (OXT) needs to be added. This atom is added so that it is in the plane of the C\alpha, C and O atoms of the residue.

Scripting usage:

(add-OXT-to-residue imol residue-number insertion-code chain-id) ⁸⁶,

where insertion-code is typically "".

Note, in order to place OXT, the N, CA, C and O atoms must be present in the residue - if (for example) the existing carbonyl oxygen atom is called “OE1” then this function will not work.

5.25 Add Atom at Pointer

By default, “Add Atom At Pointer” will pop-up a dialog from which you can choose the atom type you wish to insert ⁸⁷. Using (set-pointer-atom-is-dummy 1) you can by-pass this dialog and immediately create a dummy atom at the pointer position. Use an argument of 0 to revert to using the atom type selection pop-up on a button press.

The atoms are added to a new molecule called “Pointer Atoms”. They should be saved and merged with your coordinates outside of Coot.

5.26 Place Helix

The idea is to place a helix more or less “here” (the screen centre) by fitting to the electron density map. The algorithm is straightforward. First we move to the local centre of density, then examine the density for characteristic directions and fit ideal helices (of length 20 residues) to these directions. The helix is then extended if possible (by checking the fit to the map of residues added in ideal helix conformation) and chopped back if not. If the fit is successful, the helix is created in a new molecule called “Helix”. If the fit is not successful, there is instead a message added to the status bar. You can build the majority of a helical protein in a few minutes using this method (you will of course have to assemble the helices and assign residue numbers and sequence later).

This is available as a scripting function (place-helix-here) and in the GUI (in the “Other Modelling Tools” dialog).

5.27 Building Ideal DNA and RNA

The interface to building ideal polynucleotides can be found by pressing the “Ideal RNA/DNA…” button on the “Other Modelling Tools” dialog.

For a given sequence, a choice of DNA or RNA, A or B form, single or double stranded is presented.

The interface may not gracefully handle uracils in DNA, thymines in RNA or B form RNA.

The ideal B-form DNA is somewhat under-wound, needing 11 base-pairs to repeat (instead of the expected 10.5). There is no easy fix for this currently.

5.28 Merge Molecules

The dialog for this opperation can be found under “Calculate” in the main menubar. This is typically used to add molecule fragments or residues that are in one molecule to the “working” coordinates ⁸⁸.

The scripting interface is used like this

(merge-molecules molecule-list target-molecule)

e.g.

(merge-molecules (list 1 2 ) 0)

merges molecules 1 and 2 into molecule 0.

5.29 Temperature Factor for New Atoms

The default temperature factor for new atoms is 30.0. This can be changed by the following

(set-default-temperature-factor-for-new-atoms 50.0)

5.30 Applying NCS Edits

Let’s imagine that you have 3-fold NCS. You have molecule “A” as your master molecule and you make edits to that molecule. Now you want to apply the edits that you made to “A” (the NCS master chain ID) to the “B” and “C” molecules (i.e. you want the “B” and “C” molecules to be rotated/translated versions of the “A” molecule). How is that done?

There are now guis to NCS command to help you out (under Extensions). However, for completeness here are the scripting versions:

(copy-from-ncs-master-to-others imol master-chain-id)

If you have only a range of residues, rather than a whole chain to replace:

(copy-residue-range-from-ncs-master-to-others imol master-chain-id start-resno end-resno)

e.g.

(copy-residue-range-from-ncs-master-to-others 0 "A" 1 5)

If you want to copy a residue range to a specific chain, or specific list of chains (rather than all NCS peer chains) then make a list of the chain-ids that you wish replaced:

(copy-residue-range-from-ncs-master-to-chains 0 "A" 1 5 (list "C"))

in this case, just the residues in the "C" chain is replaced.

5.31 Running Refmac

Use the “Run Refmac...” button to select the dataset and the coordinates on which you would like to run Refmac. Note that here Coot only allows the use of datasets which has Refmac parameters set as the MTZ file was read. By default, Coot displays the new coordinates and the new map generated from refmac’s output MTZ file. Optionally, you can also display the difference map.

You can add extra parameters (data lines) to refmac’s input by storing them in a file called refmac-extra-params in the directory in which you started coot.

You can also provide extra/replacement parameters for refmac by setting the variable refmac-extra-params to a list of strings, for example:

(set! refmac-extra-params (list "REFINE MATRIX 0.1" "MAKE HYDROGENS NO"))

Coot “blocks” ⁸⁹ until Refmac has terminated ⁹⁰.

The default refmac executable is refmac5 it is presumed to be in the path. If you don’t want this, it can be overridden using a re-definition either at the scripting interface or in one’s ~/.coot file e.g.:

• (define refmac-exec "/e/refmac-new/bin/refmac5.6.3")

After running refmac several times, you may find that you prefer if the new map that refmac creates (after refmac refinement) is the same colour as the previous one (from before this refmac refinement). If so, use:

(set-keep-map-colour-after-refmac 1)

which will swap the colours of then new and old refmac map so that the post-refmac map has the same colour as the pre-refmac map and the pre-refmac map is coloured with a different colour.

5.32 Running SHELXL

Coot can read shelx .res files and write .ins files, and thus one can refine using SHELXL in a convenient manner using the function

(shelxl-refine imol . hkl-file-name)

(the hkl-file-name is an optional argument)

e.g.

(shelxl-refine 0)

or

(shelxl-refine 0 "insulin.hkl")

In the former case, coot will presume that there is a SHELX hkl file corresponding to the res file that you read in; if there is not coot will print a warning and not try to run shelxl. In the latter case, you can specify the location of the hkl file.

After shelxl has finished, coot will automatically read in the resulting res coordinates, the fcf file, convert the data to mmCIF format and read that, which generates a \sigma_A map and a difference map.

Coot creates a time stamped ins file and a time-stamped sym-link to the hkl file in the coot-shelxl directory.

Please note that the output ins file will not be particularly useful (and thus shelxl will fail) if the input file was not in SHELX ins format.

There is a GUI for this operation under the “Extensions” menu item.

5.33 Clear Pending Picks

Sometimes one can click on a button ⁹¹ unintentionally. This button is there for such a case. It clears the expectation of an atom pick. This works not only for modelling functions, but also geometry functions (such as Distance and Angle).

5.34 Delete

Single atoms or residues can be deleted from the molecule using “Delete…” from the “Model/Fit/Refine”dialog. Pressing this button results in a new dialog, with the options of “Residue” (the default), “Atom” and “Hydrogen Atoms”. Now click on an atom in the graphics - the deleted object will be the whole residue of the atom if “Residue” was selected and just that atom if “Atom” was selected. Note that if a residue has an alternative conformation, then “Delete Residue” will delete only the conformation that matches that alternative conformation specifier of the clicked atom.

Only waters are deletable if the "Water" check button is active and waters are not deletable if the "Residue/Monomer" check button is active. This is to reduce mis-clicking.

To rotate the view when in “Delete Mode”, use Ctrl left-mouse.

If you want to delete multiple items you can use check the “Keep Delete Active” check-button on this dialog This will will keep the dialog open, ready for deletion of next item.

An atom can be delete using the scripting

(delete-atom imol chain-id residue-no ins-code atom-name alt-conf)

Residues can be deleted using the scripting

(delete-residue imol chain-id residue-no ins-code)

Residue ranges can be deleted using the scripting

(delete-residue-range imol chain-id residue-no-start residue-no-end)

Chains can be deleted using the scripting

(delete-chain imol chain-id)

Sidechains can be deleted from a region:

(delete-side-chain-range imol chain-id residue-no-start residue-no-end)

or for a chain

(delete-sidechains-for-chain imol chain-id)

5.35 Sequence Assignment

You can assign a (FASTA format) sequence to a molecule using:

(assign-fasta-sequence imol chain-id fasta-seq)

This function has been provided as a precursor to functions that will (as automatically as possible) mutate your current coordinates to one that has the desired sequence. It will be used in automatic side-chain assignment (at some stage in the future).

5.36 Building Links and Loops

Coot can build missing linking regions or loops ⁹². The function can be found under Calculate -> Fit Loop -> Fit Loop by Rama Search or the scripting function:

(fit-gap imol chain-id start-resno stop-resno)

and

(fit-gap imol chain-id start-resno stop-resno sequence)

the second form will also mutate and try to rotamer fit the provided sequence.

Example usage: let’s say for molecule number 0 in chain "A" we have residues up to 56 and then a gap after which we have residues 62 and beyond:

(fit-gap 0 "A" 57 61 "TYPWS")

5.37 Fill Partial Residues

After molecular replacement, the residues of your protein could well have the correct sequence but be chopped back to CG or CB atoms. There is a function to fill such partially-filled residues:

(fill-partial-residues imol)

This identifies residues with missing atoms, then fills them and does a rotamer fit and real-space refinement.

If you want to fill the side chain of just one residue

(fill-partial-residue imol chain-id res-no ins-code)

this does a auto-fit-best-rotamer and a refinement on the resulting side-chain position.

5.38 Changing Chain IDs

You can change the chain ids of chains using Calculate -> Change Chain IDs…. Coot will block an attempt to change the whole of a chain and the target chain id already exists in the molecule.

If you use the "Residue Range" option then you can insert residues with non-conflicting residue number into pre-existing chains.

5.39 Setting Occupancies

As well as the editing “Residue Info” to change occupancies of individual atoms, one can use a scripting function to change occupancies of a whole residue range:

• (zero-occupancy-residue-range imol chain-id resno-start resno-last)

example usage:

(zero-occupancy-residue-range 0 "A" 23 28)

This is often useful to zero out a questionable loop before submitting for refinement. After refinement (with refmac) there should be relatively unbiased density in the resulting 2Fo-Fc-style and difference maps.

Similarly there is a function to reverse this operation:

• (fill-occupancy-residue-range imol chain-id resno-start resno-last)

5.40 Fix Nomenclature Errors

Currently this is available only in scripting form:

(fix-nomenclature-errors imol)

This will fix atoms nomenclature problems in molecule number imol according to the same criteria as WATCHECK ⁹³ e.g. Chi-2 for Phe, Tyr, Asp, and Glu should be between -90 and 90 degrees. Note that Val and Leu nomenclature errors are also corrected.

5.41 Rotamer Fix Whole Protein

There is an experimental scripting function

(fit-protein imol)

which does a auto-fit rotamer and Real Space Refinement for each residue. The graphics follow the refinement.

5.42 Refine All Waters

All the waters in a model can be refined (that is, moved to the local density peak) using

(fit-waters imol)

This is a non-interactive function (the waters are moved without user intervention).

5.43 Moving Molecules/Ligands

Often you want to move a ligand (or some such) from wherever it was read in to the position of interest in your molecule (i.e. the current view centre). There is a GUI to do this: Calculate -> Move Molecule Here.

There are scripting functions available for this sort of thing:

(molecule-centre imol)

will tell you the molecule centre of the imolth molecule.

(translate-molecule-by imol x-shift y-shift z-shift)

will translate all the atoms in molecule imol by the given amount (in Ångströms).

(move-molecule-to-screen-centre imol)

will move the imolth molecule to the current centre of the screen (sometimes useful for imported ligands). Note that this moves the atoms of the molecule - not just the view of the molecule.

5.44 Modifying the Labels on the Model/Fit/Refine dialog

If you don’t like the labels "Rotate/Translate Zone" or "Place Atom at Pointer" and rather they said something else, you can change the button names using:

(set-model-fit-refine-rotate-translate-zone-label "Move Zone")

and

(set-model-fit-refine-place-atom-at-pointer "Add Atom")

6 Map-Related Features

6.1 Maps in General

Maps are “infinite,” not limited to pre-calculated volume (the “Everywhere You Click - There Is Electron Density” (EYC-TIED) paradigm) symmetry-related electron density is generated automatically. Maps are easily re-contoured. Simply use the scroll wheel on you mouse to alter the contour level (or -/+ on the keyboard).

Maps follow the molecule. As you recentre or move about the crystal, the map quickly follows. If your computer is not up to re-contouring all the maps for every frame, then use Draw -> Dragged Map… to turn off this feature.

6.1.1 Map Reading Bug

Unfortunately, there is a bug in map-reading. If the map is not a bona-fide CCP4 map ⁹⁴, then coot will crash. Sorry. A fix is in the works but “it’s complicated”. That’s why maps are limited to the extension ".ext" and ".map", to make it less likely a non-CCP4 map is read.