Fitting coordinates to an SIR, MIR, or MAD map can be difficult. X-POWERFIT, X- AUTOFIT and X-BUILD (together called X-FIT) are QUANTA applications that speed up and enhance the process of de novo map interpretation, as well as the general model building in later stages of macromolecular refinement.
X-POWERFIT, X-AUTOFIT and X- BUILD capabilities include:
X-AUTOFIT permits generation of maps on the fly from within the X-AUTOFIT X-BUILD main palette. After selecting the reflection data file and generating the first map, you can add atoms/residues or rebuild the existing molecule (psf file). You can then generate a new map using CNX with a single click.
X-AUTOFIT permits up to six maps and, for each map, seven contour levels to be active at one time. It uses one of the currently open maps as the basis for the bones and real-space calculations. Bones are calculated from electron density maps.
The skeletonization algorithm used in X-AUTOFIT is based on the original rules described by Greer (1974). The four rules were modified and the algorithm reimplemented to incorporate improvements in the speed of calculation and the quality of the resulting bones. Mainchain and sidechain bones are determined by analysis, and a spline function smooths the bones segments to improve interpretability.
X-AUTOFIT generates solvent masks from coordinate and bones data. The solvent mask facility uses fast algorithms to determine solvent boundaries from coordinate information or bones (and, hence, map) information. The solvent mask can be interactively edited with a spherical pointer, and voids within the mask can be automatically deleted with a single tool.
The 3D text editor can be used to place annotations throughout the macromolecular structure. After notations are created, you can select a notation from a list and the display re-centers on the associated point in the macromolecular structure. You can also load information about the macromolecular structure into the note text utility, thereby allowing you to rapidly find problem areas during crystallographic model building.
X-POWERFIT provides an algorithm to determine the Ca trace directly from the electron density for high resolution data. For lower resolution (4.0 Å - 2.0 Å) data, there is an algorithm to determine the secondary structure from the electron density, to automatically place Ca atoms into these parts of the density, and then to extend the Ca trace until the model is close to complete. There are also algorithms for Ca refinement.
X-AUTOFIT also provides semi-automated Ca tracing and manipulation. The alpha-carbon (Ca) building facility in X-AUTOFIT intelligently places Ca coordinates into the electron density map using a rule-based process. It also allows cut-and-paste of fragments and manual editing of Ca atoms. X-POWERFIT can take the current Ca trace and search a set of pdb files for matching geometry.
Once the Ca tracing is complete, you can automatically assign the sequence to the Ca trace in cases where there is high resolution data and there are aromatic residues in the structure. In cases where the automated assignment does not work, or there are no aromatic residues in the structure, you can also carry out a fuzzy sequence assignment (such as: big, aromatic) for each residue of the Ca fragment(s). There is also an option to explicitly assign one of the 20 amino acids to a Ca atom. QUANTA uses this to show a weighted forward and backward alignment to the sequence. A secondary structure prediction from sequence tool returns the secondary structure prediction of a sequence and colors the sequence view on screen using red for helices and blue for strands.
X-AUTOFIT can create an all-atom representation using refinement techniques, database fragment fitting, and direct correlation of the Ca conformations to all-atom models. These three techniques can be used within X-AUTOFIT to fit the atoms of a residue to the map from just the Ca positions of the Ca trace fragment. The quality of fit is reported by color coding atoms in the fitted segment.
Model building is carried out with the aid of real-space refinement, regularization, and rigid-body refinement algorithms, as well as by traditional manual editing. An automated tool (Structure | Auto build) that runs through all build residues and fits them using a new mixed grid and gradient protocol. The automated tools of X-BUILD generally give a 200-to-500-fold decrease in time for a model building session and often result in improvement in the subsequent refinement of coordinates compared to traditional manual model building.
X-AUTOFIT supports three refinement techniques. Single residues can be fitted by grid searching about torsions, torsion angle real-space gradient refinement, and Monte Carlo fitting. X-AUTOFIT is designed to support the following special cases associated with electron density proteins fitting:
X-BUILD supports two kinds of validation tools designed for the crystallographic process.
The first is an entirely automated system where common errors associated with model building are detected and can be automatically fixed.
The second method of validation provides a set of functions that can derive data from molecular coordinates, apply further functions to them, and then plot them in a graph window. The molecule display, graph display, and tables are integrated, so that the table and graph can be picked to update the molecular display.
All X-BUILD tools are logged automatically in a table of previous commands, which can be used to create a log book, undo/redo any edit, assess the use of the application, and even recover and analyze changes from previous model-building sessions.
An automatic save of the current state is made every 5 minutes when in the X-POWERFIT, X-AUTOFIT or X-BUILD modules. This is in addition to the facility for user-directed saves.
Implementation of the X-AUTOFIT tools in QUANTA provides an integrated environment in which many associated tools in Protein Design, Protein Health, and Conformational Search can be combined with features of the X-AUTOFIT application to enhance the model-building and refinement process.
The rest of this chapter and the four that follow describe X-AUTOFIT, X-BUILD, and X-POWERFIT and their operation within QUANTA. These sections explain:
QUANTA has been designed to handle maps of up to 500 grid points in the x, y, or z dimensions. This size represents a maximum brick map size of 125 MB or, when a map is stored as real numbers (as in a CCP4 map), 500 MB, which is a very large map. The size of map that can be handled in X-AUTOFIT is defined by the virtual swap space of the system running QUANTA, since all map information and derivatives of these are dynamically allocated as required. A virtual swap space of at least 300 MB (preferably 400 MB) is best for the use of X-AUTOFIT/X-BUILD. This amount of virtual swap space allows the use of maps of up to 20 million grid points and any bones and mask information derived from this map. Of this space, QUANTA will use about 55 MB and the operating system approximately 30 MB. The remaining space can be allocated by the X-AUTOFIT application or used by other programs. If this happens, then you should leave QUANTA after saving any changes and restart the program. This returns all the swap space allocated for map manipulation to the system. It is not enough to just restart QUANTA from the File menu, you must exit the program.
You can display a pie chart of memory usage. There is no interface option for this, so you need to edit the xfit.pack file. The line containing "MemS 0" should be changed to "MemS 1" while the X-AUTOFIT:X-BUILD is closed. Upon entering the X-Ray application, a pie chart shows current memory usage on the computer. When available memory drops below 50Mbytes the pie chart turns from green to orange, and when available memory drops below 20Mbytes the pie chart turns red.
A real memory size of at least 48 MB is recommended for using X-AUTOFIT. If large structures are to be studied, 64 MB provides a more responsive program.
MEM_Alloc: Allocation error [-27316]
===============================
A request for memory was made:-
memory = 27 Mbytes
But the system refused this request
Either : Increase system swap space
Use a small map
Finish other programs running
===============================
Graphical objects generated under SGI GL and Open GL cannot have their memory returned to the system after use. This means that continued re-display of different areas of the map gradually uses the available swap space when very large maps are used. X-AUTOFIT therefore continually monitors the available swap space and prints warning messages to the textport if the available space drops below 10 MB.
==========================
You are strongly advised
to save and exit QUANTA as
swap memory is low
The current OS does not allow
this to be fixed
==========================
Since it is impossible to prevent a program from being killed by the system when all the available swap space is used, you should take one of the following actions if this message appears:
a. Close other programs on the system to release memory.
b. Delete bones and/or delete a map mask in X-AUTOFIT.
c. Close the current map and use a smaller map.
d. Finish editing, save all changes in X-AUTOFIT, and exit QUANTA.
The memory tests in X-AUTOFIT help prevent crashing due to memory depletion.
This section describes how to access and get started with X-AUTOFIT.
Prior to starting X-AUTOFIT, do the following:
For learning purposes, start with a 2fo-fc electron density map with a corresponding set of coordinates. You may have to convert an existing map using the map conversion facility on the Maps Management table and palette. For more information, see Managing Maps.
X-AUTOFIT is accessed from the Applications menu in the QUANTA main menu, as follows:
1. Create a new directory in which to run X-AUTOFIT. Move to that directory and start QUANTA.
2. Select X-AUTOFIT from the Applications menu. The main X-AUTOFIT:X-BUILD palette and the Pointer palette appear, as well as a graph window containing the allowed regions for a Ramachandran map and any points representing the phi-psi angles for any known coordinates. If the map table is not open, then open it using the Map show table (select Map table | Show map table from the Draw menu) tool.
An Object Management table also opens. As objects are generated, they are added to the table: bones, mask, symmetry atoms, Ca trace, and 3D text.
The Object Management table can be used to toggle the relevant information on and off. Objects can be deleted from this table, but X-AUTOFIT generates them again if required.
The following description indicates the aims behind X-AUTOFIT and generalizations about its use.
X-AUTOFIT allows only one structure to be edited at any time, to prevent possible confusion and ambiguity. The current editable molecule is the first active and displayed molecule in the object management table. To change the molecule that you are editing, use the Molecule Management table to change the activity of the molecule.
QUANTA can use six active and contoured maps at the same time. Because X-AUTOFIT uses crystallographic information for real-space refinement and bones, you must identify which map to use if more than one map is open. To change the current map, use the tool X-AUTOFIT | Change RSR-bones map.
By default, if any maps are open when you first enter X-AUTOFIT, the first map that you open is the map used for RSR and bones calculations.
X-AUTOFIT differs slightly in the use and behavior of palettes from the rest of QUANTA. Twelve palettes can be open in X-AUTOFIT, and all can be open and used simultaneously. This allows almost any type of manipulation and calculation necessary for macromolecular crystallography, while simplifying the number of commands visible at one time.
All tools in the X-AUTOFIT application are active and can be used from any of the palettes at any time, with these restrictions:
1. When an editing process is active, (when the Accept or Quit dialog box is present), Accept or Quit are the only options you can choose.
2. X-SOLVATE and X-LIGAND are separate applications and do not allow other X-AUTOFIT tools to be used while they are active. Because they are automated procedures, they take control of screen placement and map display.
The main X-AUTOFIT:X-BUILD palette has ten sub-palettes, the X-AUTOFIT | Sequence palette has two sub-palettes, and the X-AUTOFIT | Build palette has two sub-palettes. The Accept or Quit dialog box appears when the application requires you to choose whether to accept or abort the current editing or picking process.
The general use of X-AUTOFIT can be divided by activity. The following table indicates the palettes required for a given operation. You can have more palettes open at a time.
|
X-AUTOFIT | Build atoms | Color atoms |
|
A user-defined palette is also provided that allows various tools from different palettes to be put in one place. This also allows you to use macros.
The Text and Pointer palettes can be used at any time, and normally the Pointer palette should be open at all times. Each palette is made active by selecting the tool on the parent palette and made inactive by selecting Hide this menu at the bottom of each palette. If a palette is active then the parent tool is highlighted. If the parent tool is highlighted but the associated palette is covered by another window, selecting the highlighted tool for that palette brings the palette to the front.
The X-AUTOFIT:X-BUILD main palette contains three other tools, for saving the results of the Ca-tracing, saving the results of the model building, and restoring the old MSF from disk. When you exit X-AUTOFIT, it writes a Ca trace session file to disk. It also writes this file when the X-AUTOFIT | CA Build | Save changes tool is used. This file also contains the sequence for the alignment and any sequence assignment made to Ca atoms. The data is read back into X-AUTOFIT the next time the program is used. X-AUTOFIT also checks to see if any changes have been made to any molecular coordinates, and a dialog box asks you to save each changed molecule before leaving X-AUTOFIT.
When you exit X-AUTOFIT, it stores the current status of parameters and settings and the current open palettes, so that a new session starts from where you left off.
There are four default sets of dials in X-AUTOFIT. While using the edit functions, the dial set appropriate to the requested function is displayed. Because some functions can use more than one dial set, there are options on the Pointer, Mask, and CA Build palettes to set the type of dial in use.
Two of the default dial sets are used to adjust the position of Ca atoms while fitting Ca atoms to MIR/SIR or MAD density. These control the position of the current active Ca atom.
The third dial set controls the placement of a cursor on the screen. This cursor is normally used for all other building procedures not involving Ca-atom tracing into density. The selection of the dial set also controls the type of plot displayed in the graphs window, as described in the next section.
The fourth dial set controls the mask pointer. This set allows you to position a spherical pointer and the adjust its radius. It becomes active after the calculation or reading of a mask. The X-AUTOFIT | Mask | Mask dials tool changes the dials to the mask dial set, if a mask is currently active.
The X-AUTOFIT | Pointer | Pointer-dials tool changes the dial set to the pointer dials, and the X-AUTOFIT | CA Build | CA dials tool changes the dials to one of the Ca-building dial sets relevant to the current active Ca atom.
The default X-AUTOFIT dial set fills the U dials and by default is always displayed while using X-AUTOFIT, X-POWERFIT, or X-BUILD. A toggle on the X-AUTOFIT:X-BUILD | Options... dialog box can be used to reset the default dials to that of the main "1" dial set of QUANTA if preferred.
Two graph windows are generated in X-AUTOFIT. The first is always present in the lower right corner of the screen. This graph window displays three sets of information.
The second graph window is generated as part of the advanced validation tools and can display many different types of data from the validation tables. Details on the use of this graph, generation, annotation, and plotting can be found in Advanced validation techniques. The tools are described in more detail in Tables and Graphs.
Four tables can be created from X-AUTOFIT | X-BUILD. The first is the last-command table. The last-command table is opened from the main X-AUTOFIT | X-BUILD palette and provides a catalogue of tools used, time and date, undo/redo option, and user-defined remarks. The menu items at the top of this table allow various actions to be carried out based on the last-command table
The remaining three tables are generated while using the advanced validation tools. One table is generated for any property or function based on atom data, the second is generated for residue data, and the third when data is produced that has no direct relation with coordinate information. The atom and residue tables can be picked to place the molecular view and selected (by row or column) for the various manipulations and plotting. Use of the tables can be found in the section on Advanced validation techniques. The tools are described in more detail in the section on Tables and Graphs.