The model building functionality in X-AUTOFIT contains many features that accelerate the modeling of crystallographic coordinates. Model building allows real-space refinement in torsion angle space, rigid-body refinement, regularization, and general manual editing of macromolecular coordinates. Auto Build allows you to walk through the structure and rebuild the main chain and side chain of each residue in an automated manner. The Add/Delete palette allows the addition of amino acids, waters, and ions, deletion of residues and ranges of residues, and addition of alternative conformations to amino acids.
To avoid confusion about the target of issued commands, only one structure can be edited at a time. To define the structure to be edited in X-AUTOFIT, use the "active" property on the molecule management palette. The first active and displayed molecule in the molecule management table is taken as the molecule to edit.
X-BUILD supports the editing of polypeptide and polynucleotide structures. Since the basic structure of these macromolecules is different, the tools described in this section behave differently depending on the residue selected. The program determines from the residue selected which functionality to use and which information to show in the plot window.
Protein atom naming is very strictly observed, and the conventions used are those for all the major crystallographic programs. The 20 basic amino acids and some modified amino acids (that is, HISH) are recognized by the program.
X-BUILD automatically supports both ribose and deoxyribose units when editing nucleic acids, and any addition of residues is carried out using a sugar of the same type as the proceeding residue in the chain. Currently, only the five standard bases, adenine, cytosine, guanine, thymine, and uracil, are supported for full editing, and no restriction is made on the uracil/thymine pair for RNA and DNA. The 5'-phosphate group is optional and is not required for editing the terminal residue in a chain, but dummy atoms are added during regularization and refinement and then removed on completion of these functions. Also supported are the 5'-terminal patches of no-phosphate, phosphate + no terminating oxygen, and phosphate + terminating oxygen.
The atom- and residue-naming convention for DNA and RNA tends to be less strict than for proteins. The PDB standardization has been used for X-BUILD: 3-letter residue names, and the use of the quote and not * for sugar atom naming. To make the program more general, entry into X-BUILD automatically changes the residue names and atoms names if these conventions are not used in the original file. The phosphate atoms are called "P", "O1P", "O2P", and the terminating 5'-oxygen "O5T". The terminating hydrogen atoms are "H3T" and "H5PT" where applicable. Some of the editing facilities in X-BUILD are not available if these naming conventions are not used. (The same conventions are used within CNX, CHARMm, and CCP4).
Both protein and nucleic acids can be edited in nonhydrogen, polar-hydrogen, and all-hydrogen modes. As with proteins, the program automatically detects the hydrogen mode on entry to X-BUILD, although you may override the choice at any time.
Disorder, nonbonds, energies, and other general functionality with X-BUILD is supported for both polymer types.
The number of atoms displayed around the specified center is defined by the tool X-AUTOFIT | Options... | Coordinate radius. The volume of map displayed is defined by X-AUTOFIT | Options... | Map-radius, while the number of symmetry atoms is defined by the tool X-AUTOFIT | Options... | Symmetry-radius. The size of the calculated bones region is defined by the map radius.
X-BUILD has the notion of a current residue. The initial value of the current residue is "1", but this is changed by several commands. Its value is restricted to between 1 and the total number of residues in the QUANTA data structure. It can be set but not used if it lies outside this range. The actual value therefore refers to the sequential position of the residue in the entire QUANTA data structure. The value of the current residue is saved between sessions. The value is set by the placement of view commands and used to color the current residue point within the Ramachandran and DNA circle plots red, which provides feedback information between these plots and the molecule display.
You can switch the program to amodal/active residue mode, in which the current residue is highlighted and any tool acts immediately on this residue. (A residue range cannot be handled in this way.) This mode of action can be used to make single residue editing quicker, since less mouse-clicking is required.
When building all-atom models, X-BUILD can act in both modal and amodal fashion. This is controlled by the Active residue on tool in the Pointer palette.
When in amodal mode, the current residue is labeled with pink rhomboids. The current residue can be changed by picking another visible atom. Tools are provided to regularize a range of residues from a single residue, and a tool to regularize a volume of residues from a single residue - these permit the use of the modal action on multiple residues.
The center of the display can be updated using several methods. The application moves the screen center, updates the map contouring around the region required, recalculates the symmetry atoms, and, if bones are active, recalculates them for the current map region.
Placement by atom coordinate: The tool X-AUTOFIT | Pointer... | Place using coordinates prompts you to select a single atom position from the main screen display. Selecting an atom (including a symmetry atom) results in the display center moving to this coordinate. Selecting a non-atom position on the screen or selecting any palette tool aborts this operation. This tool sets the current residue pointer to the residue number of the picked atom, and the Ramachandran plot or DNA plot is updated.
Placement by bones: If the bones are active, the tool X-AUTOFIT | Pointer | Place-by-bones prompts you to select a bones point from the main display. Selecting a non-bones position on the screen or selecting any palette tool aborts this operation.
Placement by atom name: The tool X-AUTOFIT | Pointer | Place-by-atom opens a dialog box that allows specification of an atom name, a sequence ID, and a segment name. If this atom is present in the currently active and displayed molecule, it becomes the new display origin. If it is not found, no change occurs. This tool sets the current residue to that of the residue number of the named atom, and the Ramachandran plot or DNA plot is updated.
Place at next residue: The tool X-AUTOFIT | Pointer | Place-at-next-residue moves the display center to the next residue. If the next residue is an amino acid, the display is centered on the Ca atom. If the next residue is a water or an ion, then that residue becomes the display center. If the next residue is neither an amino acid nor a single-atom residue, the display center is set to the first atom in the residue. This increments the current residue number by one. (You can change the next-residue step to values greater than 1 in the Options... dialog box
Place at previous residue: The tool X-AUTOFIT | Pointer | Place-at-previous-residue moves the display center to the previous residue. If the previous residue is an amino acid, the display is centered on the Ca atom. If the previous residue is a water or an ion, that residue becomes the display center. If the previous residue is neither an amino acid nor a single-atom residue, the display center is set to the first atom in the residue. This decrements the current residue number by one. (You can change the previous-residue step to values greater than 1 on the Options... dialog box
Placement by Ramachandran plot: The graph window is always active in X-AUTOFIT and shows a Ramachandran plot of the current active and displayed molecule while the Ca placement dials are inactive. Green points on this plot indicate non-glycine residues, and blue points indicate glycine residues. The red point indicates the current residue. To place the screen origin, click any point in this Ramachandran plot; the residue that has the phi/psi values for this point becomes the screen origin, and the information for this residue is written to the textport. The point picked turns red to show that it is the current residue. There is an X-AUTOFIT Options | Center at Ramachandran point toggle that allows a Ramachandran point to be selected without the display center being updated, to allow the Ramachandran plot to be interrogated without waiting for the screen center to be updated. This X-AUTOFIT | Options option must be turned on for the display centering function to be effective.
Placement by DNA circle plot. This plot is available when DNA or RNA coordinates are edited. Centering using this plot is identical to that for Ramachandran plots.
The validation graph is generated from the Tables and Graphs functionality in X-BUILD. If any point in the graph is picked, then the residue/atom that was used to generate this point becomes the center of the display. If the graph plot was generated from general data (i.e., does not relate to a atom or residue), then no update is made. All maps and bones and additional properties are updated by this pick.
If the atom or residue table is open from the Validation functionality, then this can be used to update the display. If a table row is picked, the atom/residue within the current molecule that relates to the row of data in the table becomes the center of the display. The map and bones are not updated by this pick. To update this information: after picking the validation table, use the Goto pointer tool from the Pointer palette.
The Last Command table can be used to center the display at the residue defined in the residue field for the command card. If a residue range is shown in the residue card of this table entry, the display is centered at the first residue of the range.
Placement by text marker: The screen origin can be updated using 3D text marks defined with the Text palette. To go to the next text in the list of text marks, use the tool X-AUTOFIT | Text | Next-text; go back to the previous text with X-AUTOFIT | Text | Previous-text. The current loaded text marks (that is, user defined or loaded property) can be used with the next and previous tools; the most obvious text mark for this process is "loaded property = water molecules". To go to a particular text mark, use X-AUTOFIT | Text | Defined-text. A scrolling list of the current text marks appears, from which you can select a single text. This tool does not change the current residue number.
A Ramachandran plot appears after any edit of an amino acid and remains until any subsequent edit of a nucleic acid or Ca trace atom. The Ramachandran plot drawn in X-AUTOFIT is a hard-sphere contact and 10% overlap surface of the angles f (Ci-1-Ni-Cai-Ci) and y (Ni-Cai-Ci-Ni+1) (Ramachandran 1968). The contour lines shown are taken from Dickerson and Guis. The definition of the Ramachandran angles are described in Ramachandran and Sasisetharam (1968).
When you are not editing any part of the structure under X-AUTOFIT | Build atoms, the Ramachandran plot shows the current values of f and y for all the residues in the active and displayed protein as colored points on the Ramachandran plot. All non-glycine residues are green and all glycine residues are blue. The current residue is drawn with a red circle.
The number of points drawn on the Ramachandran plot can be changed using the X-AUTOFIT:X-BUILD | Options... dialog setting:
[X] Center at Ramachandran point...
Segid : [ * ] start : [ 1 ] End : [ 9999 ]
This option allows the specification of a range of residues to be drawn on the Ramachandran plot, enabling you to limit the amount of information presented in this plot window when editing small regions of the molecule. The specification is defined for the first active and visible molecule. The wildcard specification (*) indicates that any segment name is allowed, and the start/end values are those of the sequence ID of the residues. This tool is also valid for DNA and RNA models.
To identify a point on the plot, click the point. The residue that corresponds to that point picked is shown in the textport. If the option X-AUTOFIT | Options | Goto-Rama residue is active, the display is also centered on this residue.
When editing some property under X-AUTOFIT | Build atoms, such as the backbone peptide plane (X-AUTOFIT | Build atoms | Change backbone), the current Ramachandran angles of only the affected residues are shown. In this case, when the angle Cai-Ni-Ci+1-Cai+1 is rotated, the f/y angles for residues i and i+1 are shown. For a detailed description of which point corresponds to which residue while editing the coordinates, refer to X-AUTOFIT:X-BUILD Tools for descriptions of the relevant tools.
The DNA/RNA plot window appears automatically after any edit of a DNA/RNA residue and remains until any subsequent edit of an amino acid or Ca trace atom. The plot window style is retained between editing sessions.
The DNA/RNA plot consists of seven fields, a-b-g-d-e-z-x, in decreasing concentric circles. The coordinate frame is polar, where the radial values have discrete values for each torsion in DNA residues. The polar angle is the value in degrees for each of the seven allowed torsions. Seven green circles are drawn for each residue in the polynucleotide, one in each of the seven fields of the plot. (Terminal residues may have incomplete fields where atoms are not present in the residue.) The current residue (see The current residue value on page 79) is indicated by red circles. If the plot is picked using the cross-hair cursor, then the nearest point on the plot becomes the current residue (as indicated by a red circle), the graphical display changes to place this residue at the origin (when Center at Ramachandran point is on), and information on this residue is shown in the textport.
When you are editing some property that will affect a torsion value, the fields are filled only with those values relevant for the edit.
Various kinds of information displayed in the molecular view and plot windows can be picked, although there are restrictions.
1. The atom information can be picked at any time to return an atom label for the atom picked.
If X-BUILD is in active-residue mode (amodal), then picking any atom changes the current residue. This changes all graphical objects associated with the current residue.
2. The Ca trace can be picked only when bones are not turned on. This restriction results from the use of bone-picking to define the Ca position. The Ca trace atom is labelled with the current segment number and Ca atom position in the segment. The segment number and Ca position are transient and depend on the current segment and atom (CA build | current seg-res) when the label is generated.
3. The bones can be picked, when the Ca dials are active, to define the position of a terminal Ca trace atom. Only the terminal Ca trace atom can be placed by this action. The Ca trace atom is moved so that the current position points at the bones point picked, and the pseudobond to this is retained at 3.8 Å.
4. The Ramachandran and DNA plots can be picked at any time to reset the molecular view origin, map display, and bones display to the residue picked. Information on the Ramachandran/DNA mainchain torsion angles are written to the textport with the residue information.
5. The Ca plot can be picked when active, and the current Ca atom terminates the current segment. The current Ca atom is placed in a conformation defined by the point picked.
6. Pick the pointer at any time to return the real-space coordinate of the pointer.
7. Picking an entry in the Last Command table will position the molecular view, provide details on an edit, undo/redo an edit, or provide a comment card, depending on the table cell picked.
8. Picking the validation tables for atom and residue information updates the molecular view so that the center is defined by the atom/residue row in the table. Map and bones information is not redrawn.
9. Picking the general graph updates the molecular view so that the center is defined by the atom/residue information that generated the plot. Map and bones information is also redraw.
Symmetry cards for each molecule used are written to files with the same name as the MSF file, but with a .sym suffix. This allows symmetry to be accessed quickly. The symmetry of the Ca trace and bones is stored in the Xfit.sym file, since there is no saved MSF information for this data. If the symmetry is the same for all molecules, the program uses the single set of symmetry information stored within the program. This allows the symmetry to be drawn very quickly for simple cases but also allows more complex cases of multiple symmetry to be handled.
X-BUILD now allows symmetry atoms to be picked and the information of the picked atoms is displayed.
Non-crystallographic symmetry (NCS) is handled by X-AUTOFIT, X-BUILD, X-LIGAND, and X-SOLVATE. Up to 60 NCS matrices can be entered for each open molecule and the Ca trace. The NCS symmetry is displayed in a different color from the normal symmetry information and can be picked.
A facility is provided in X-AUTOFIT-X-BUILD to run an external program. Up to 20 external programs can be associated and, when set up, appear on the Run External Program palette.
The external program facility is set up by writing scripts that define:
The scripting language supports hidden dialog boxes, variables, and various language constructs that provide a rich variety of possibilities to set up support programs.
You should provide a script as a file named script.# where # is a number between 1 and 20. When present, this file is checked and added to the Run External Program palette when this palette is opened. When this new tool is selected, the script is read and the dialog opened as defined by the script. When the dialog is completed, the command file defined by this script is written and run.
A debugger is provided that indicates problems with the script.
The tables and graphs palette is used for advanced analysis of proteins (in particular) and other macromolecules. The advanced validation techniques are carried out by generating tables of data based on atomistic and residue properties and applying functions to them. The data can then be plotted and the graphs used to identify features of the molecule that are interesting or in error.
This palette's tools are used to generate tables containing information on atoms, residues, and other general data. The data tables can be operated on by several functions and plotted in various styles. The graphs and tables can be picked to center the molecule view, and the graphs can be annotated and plotted to a PostScript file.
Tables of atom and residue information can be generated:
The length of the tables is defined by the number of visible active molecules in the molecule table and the atom selection defined by the atom selection tool. If two molecules are both active and visible, then the table contains the basic information for both molecules. Once created, the table length of the atom and residue tables cannot be changed; it is necessary to delete the tables to change the table length. The General table is the length of the longest set of data.
The protein property functions produce data only from the first active and visible molecule in the molecule table, all remaining values not associated with this molecule that are in the table are given a value of NoData. These values are not plotted. If you want to plot data of the same property from two molecules, then change the activity of molecules in the molecule table and re-apply the same property for the second molecule. Both data columns can then be plotted.
Difference data is generated with respect to the first current active molecule in the molecule table.
Column functions and calculations apply directly to the data columns selected and are independent of molecule visibility and activity.
Data can be picked from the tables using the following procedures:
Table 1 is the atom table and contains rows of data that correspond to the atoms in the current visible and active molecules that are selected based on the selection criteria defined by the tool X-BUILD | Tables and Graphs | Atom Selection.
The residue table contains rows of data for each residue in the current active and visible molecules. The selection of residues is based on the selection criteria defined by the X-BUILD | Tables and Graphs | Atom Selection tool.
The general table is used for data that has no direct relationship with either the atom or residue data of the current selected and visible molecules.
You cannot add data directly to the scratch table, which is used by the program to store information for automated validation.
Once data has been created in a table, no other rows of data can be added to the table via the properties, difference, calculated, or function tools. If any of these tools are used, only data that already exists in the original atom table is updated in the residue table; all other data is set to NoData (if in the residue table but not in origin data) or ignored (if in origin data and not in the residue table).
Inconsistent data may be generated if:
General graph drawing facilities are found on the Tables and Graphs palette. Graphing works in conjunction with the tables facilities described above to produce publication-quality plots and for general analysis of properties.
Three tools on the Tables and Graphs palette affect data plotting. Selected columns of any of the atom/residue/general data tables can be plotted on the same graph in different colors and styles. The number of graphs to be drawn is defined by the number of selected table columns. If no table columns are selected, three blank plotting definitions are provided for you to fill in. The plot data tool opens a dialog that provides options to define the look of a plot. This includes the color, line width, style, legends, and axis labelling.
The allowed table names requested in this dialog are Atom, Residue, and Data and specify the origin table of the data. If a column is selected, this information along with the column number is filled in by the program.
Only one x column is provided by this functionality, and all data is plotted against this one set of x data. By default, the column header for the x data is zero. A zero column number specifies that the data is incremental data, starting at 1 and incremented by 1 for each data value. Any plotted data from the residue or atom table is automatically linked to the molecule data, regardless of the x ordinate.
Legends can be added from the plotting setup palette and can also be added using the annotation facility. The font size used for the legends and the axis labels is defined by the Axis font option list, and the Axis style changes the border around the graph.
General annotation is added with the Label graph... palette, which allows you to place text, boxes, lines, and circles in various colors and styles. The text style is defined using the Label string, Size option, Orientation value, and Color value. The box, circle, and line styles are defined by the Color value.
The Attachment defines whether the annotation is to be added to the Graph data or is just an annotation in the Figure window. Most annotations, such as legends, are added as Figure annotations. If the annotation relates directly to a data point(s) (such as a data point desciption) then it should be added as a graph annotation. Graph annotations cannot be placed outside the bounding box of the graph.
To delete an annotation, select the Delete button. The application waits for you to select an annotation on the graph. Selecting the Add button requests a single point pick for a text string (the bottom left end of the text is placed at the point picked), while line, circle, and box annotations are placed by picking two points in the graph window.
The PostScript settings... dialog affects only a plot produced in a publication-quality plot style and only direct plotting of the graph window contents as a PostScript file is possible from this dialog.
The scaled font toggle is provided so that the whole plot plus any annotation can be scaled to the size of the page. When this is not checked, the font size is set as defined when the plot was generated and labelled.
The color options allow you to generate plots in black and white, color with the black and white objects (including background) reversed, and inverted color. These options are provided because the graph plot on the screen has a black background, while a hardcopy plot has a white background. Color invert often produces better looking plots than Color B<->W because of the inverting of the background.
The middle nine buttons are used to place the plot on the page. The large rectangle defines the page and the smaller rectangle within this is the graph border and does not include the axis labels. These include setting the aspect ratio, translation, and scaling. The Step defines the increment of movement of the plot and is in inches.
The set button sets the current values of the options for future use, the plot button plots the current validation graph in the current position and with the current settings, and cancel aborts all actions.
The Last Command tool opens the Last Command table, where the record of all edits done in X-BUILD is stored. The table can be used to check the progress of the model-building process, undo and redo each command, analyze the work done, and create log files of X-BUILD functionality
To hide the table, click the Last Command tool again. The presence of the table does not affect the saving of commands to the file. As each command is issued (and if you accept the changes made by that tool) it is added to the top of the table.
The application records the tools used in X-BUILD, and on normal exit (finish), the current session information file is moved to a new file named command#.stack (# = 1-999). When there is a premature exit, the current session information file remains as command000.stack. Thus, the program is made aware of an incorrect exit and previous tools can be recovered. If a new session is started after the previous session was exited correctly, the Last Command table is empty (except for a save changes always made on entry). If the previous session ended incorrectly, then the Last Command table contains the previous session entries.
For each tool used in X-BUILD, a new entry is added to the Last Command table.
(*) Nonreversible command, normally insertion/deletion of residue data.
a. N/A - The command cannot be repeated/undone (various reasons - usually due to residue deletion/insertion).
b. No - The command has not been undone/redone.
c. Yes - The command has been redone/undone.
The headers at the top of the tool entries can be picked to sort the data and analyze the content of the table.
The cells in the table can be picked to place the molecule view, determine the changes made by the tool, undo changes, and redo changes.
For amino acids, X-AUTOFIT provides two forms of real-space refinement. The first (X-AUTOFIT | Build atoms | Refine 1 residue) is a true refinement procedure that carries out a gradient minimization of a residue to the current electron density. This can take a few seconds to complete and can be aborted by clicking the screen. This algorithm can refine the position and orientation of any residue type, and if torsion angle information is present in the .GSD file or a user editable file (lig.rot), then a full torsion angle minimization can be accomplished. Refinement of multiple residues by real-space refinement torsion angle refinement is carried out with the tool X-AUTOFIT | Structure | Refine zone.
Specific to amino acids and nucleic acids are X-AUTOFIT | Build atoms | RSR sidechain and X-AUTOFIT | Build atoms | RSR main chain, which use an optimized grid-search algorithm to refine to completion amino acid coordinates in 0.1 sec. The specific amino acid refinement routine also searches all possible conformations available to the amino acid and so does not get into the nearest false minima. These specific routines are recommended for amino acid fitting to density, since they are far more powerful than the traditional gradient-refinement protocol. The specific amino acid routines are also fast enough to allow editing of atom positions in an amino acid while the refinement is active. This means that the effect of placing an atom at a position can be observed as a function of electron density fit.
Where the mainchain is not yet well fitted or a residue has a poor Ca chiral volume, the simple RSR sidechain tool sometimes seems to produce spurious results. Often, the geometry of the backbone interferes with fitting the sidechain. If this happens, you may want to pick up the Ca atom with the Move atom and RSR tool and move the atom until the sidechain snaps into the density. It is usually obvious when the sidechain has found the right density. Sometimes you only need to move the Ca atom a very small distance.
A tool X-BUILD | Build atoms | Add-delete | Add 1 torsion allows the definition or removal of a torsion angle in a residue that is not an amino acid or nucleic acid. This can then be rotated manually (Build atoms | Edit chi angles, Flip torsion 180 degrees) and also refined with any of the real-space torsion angle refinement algorithms (Build atoms | Refine 1 residue, Structure | refine zone).
Missing and incorrect atoms found in residues described in the .GSD file (for example, amino acids) can be built or removed automatically using the X-AUTOFIT | Build atoms | RSR sidechain and X-AUTOFIT | Build atoms | Geometric conformation tools. Both these routines replace a sidechain of an amino acid with a template structure that has a reasonable conformation, regardless of the starting coordinates. This includes the adding or removing of hydrogen atoms when the current X-AUTOFIT non-hydrogen/polar-hydrogen/all-hydrogen building mode is different from the MSF model. When missing or incorrect atoms are found in these template residues, X-AUTOFIT initializes the temperature factors, retypes all the atoms, and resets the charges to template values. You can use X-AUTOFIT | Build atoms | Edit residue info to access the full editable table, which allows you to change B values and occupancies for all residues of a macromolecule.
The regularization protocol corrects bond, angle, and improper-torsion parameters for all templated residues in the .GSD file. If the residue to be regularized is not found in the .GSD file, the program generates the required parameters from a PSF that is generated using the type definitions of the atoms of the residue. This allows the regularization of any set of atoms in QUANTA or X-AUTOFIT. Before regularization, you must have assigned the proper atom types to the atoms.
The regularization implementation allows atom position editing while the refinement is active. This allows regions of atoms (for example, a loop in proteins) to be manually edited, with the program retaining the correct geometry during the editing process. This powerful facility for loop editing is an improvement over the usual practice of placing the Ca atoms in specific positions while the program moves the rest of the structure to compensate for the loop changes.
Regularization changes the bonds, angles, planes, and chirality of residues. It does not explicitly change the rotatable bonds (that is, chi, phi, and psi). The torsions are affected if this is the only way the algorithm can satisfy the restraints.
RSR changes the c angles and adjusts some bond angles to help find density for sidechains. The algorithm rebuilds the residue first from the template, so after using the RSR sidechain tool, all bonds have the correct value regardless of the starting positions of the atoms. The bond angles and improper torsions may have a deviation of up to 12° from the parameter values to enable the algorithm to search density when the mainchain atoms are not in the correct positions (as judged from the density). Hence, although the regularization and RSR algorithms are complementary and change different parameters, you should use real-space refinement on a residue first and regularize it afterwards.
The tools Refine 1 residue, Fit sidechain by RSR, and Move atom +RSR on the Build atoms palette do not carry out geometry minimization and thus do not correct the geometry of a residue. This tool is specifically designed to place atoms into electron density with some detriment to angles and improper torsions, the latter being easy to correct with the Regularize tool or with general reciprocal-space refinement.
The Refine Zone tool on the structure palette uses mixed parameterization and therefore fits atoms to electron density and optimizes all geometric terms.
X-AUTOFIT has two ways of treating disulfide links during regularization. It can ignore them or treat them by including the link and the connected residue. You control this behavior by setting the parameter Regularize across disulfides on the X-AUTOFIT | Options dialog box.
When the Regularize across disulfides parameter is on, if the zone you want to regularize contains a cystine as part of a disulfide, X-AUTOFIT checks whether the other cystine of the pair is already in the regularize zone. If it is, then it just sets the restraints for the bond. If the other cystine of the pair is not in the regularize zone, then the cystine is included, the N and C atoms are fixed for this residue, and disulfide bond restraints are added. Thus, if you want to regularize a disulfide bridge, select just one of the cystine residues to give a single residue range; the other residue is added automatically.
X-AUTOFIT supports no-hydrogen, polar-hydrogen, and all-hydrogen representations of residues. The mode is automatically determined on entry to X-AUTOFIT. The X-AUTOFIT Options dialog box has an option to set the hydrogen representation mode. If the loaded MSF structure has a different hydrogen representation from that defined in X-AUTOFIT | Options, then only RSR sidechain and geometric conformation tools can be used. These tools then build or remove hydrogens from the residues. Once a residue has the correct hydrogen representation, it can be edited by all the X-AUTOFIT | Build tools. Any residue not in the .GSD template file can be edited regardless of the hydrogen representation, since no atom editing on these residues can take place in X-AUTOFIT.
X-AUTOFIT supports disorder for all residues, up to four alternative positions of atoms. For a disordered residue, most commands first split the residue into multiple complete residues, depending on which conformation was selected.
If the atom selected for any tool is in a disordered residue, then:
Since most tools act on a residue (and not on single atoms), it would be normal to explicitly select the conformation to edit by selecting the required conformation from sidechain-disordered atoms. After editing a disordered residue, the separation of the atoms from the multiple conformations is checked. If the separation is less than 0.01 Å, then this atom group becomes a single atom. Otherwise, the multiple atom positions are retained as separate conformers.
The clamping option is found on the X-AUTOFIT | Options dialog box. When clamping is turned on, X-AUTOFIT moves the B (C and D if relevant) conformer so that the mainchain atoms are aligned with the A conformation mainchain atoms.This means that when a residue A conformation is edited (for example, regularized), then on completion of the edit, the B (C and D) conformers move to the equivalent before-edit position, relative to the mainchain atoms of the A chain. You should use clamping of the B (C and D) conformer to the A conformer for all single-residue disorder, and only turn off this option when a disordered loop is being added.
It is possible to convert a multiple conformation back to a single conformation by fitting the disordered atoms in the same place. When the edit is complete, a routine removes all the excess atoms from the alternative conformation, and it is possible to remove all of them if they are very similar in conformation. The residue is therefore no longer classed as an alternative conformation. The most likely scenario is when multiple disordered conformers are both fitted with RSR sidechains: this causes the same solution to be found. B conformations are fitted when added with the Add alternate conformation tool, so a B conformation can be refitted by deleting it and adding it again.
C and D conformations cannot be fitted automatically to density, though the Refine 1 residue tool may fit these if close enough to a solution. Be aware that if the mainchain atoms move for a clamped residue, then this edit may affect other disorder of this residue. (You should fix the mainchain atoms for a clamped residue before using this tool.)
X-AUTOFIT supports various naming conventions for C-terminal COO groups. X-AUTOFIT does not explicitly use or show the second oxygen position during editing, but adds it back to the residue on completion of the edit.
The tool to repatch the 5'-terminal in DNA supports three conventions. The list of options is found in Using X-AUTOFIT. X-AUTOFIT tools handle all forms of the patch, but when there is no 5'-phosphate present, these are added as dummy atoms during regularization and refinement and then removed afterwards.
For most residues, it is often only necessary to RSR the sidechain coordinates, followed by some regularization. This makes model building a very easy process. When you need to more extensively edit sections of a chain trace, use the tool Move zone and regularize. During this process, the Ca atoms of the zone being regularized and edited should be moved to the region of density that represents the Ca density. This should drag the rest of the residue atoms in the zone to a new location. Once this Ca atom movement has been completed, use the tools on the X-AUTOFIT | Build atoms palette, RSR sidechain and RSR main chain, to automatically place the other atoms.
Where the sidechain positions are indeterminate, the tool X-AUTOFIT | Build atoms | Geometrical conformation can generate a theoretical conformation based on the conformation observed from the protein databank. Each conformation is shown with the nonbond contacts to other displayed atoms.
Gradient body refinement can be used on waters to place them into the density after refinement. The process of refinement often shifts waters out of the center of the density. This tool simply returns them to a refined position. Used in conjunction with Goto next residue, this allows a simple way to traverse all the waters. This procedure is fully automated under X-AUTOFIT | Structure | Do all.
The X-AUTOFIT | Build atoms | Color atoms palette allows easy checking of a poor structure visually. After refinement, Color by B-value allows visualization of the badly fitted regions, and Color by fit shows the current quality of fit to the map. Color by progress allows the visualization of the modeling session, so that it is easy to find how much has been fitted. Also it can be used to see when the last save to disk occurred and whether a failure to fit a region would be worth undoing from disk or editing again.
The manual editing facilities provide all the necessary facilities for fitting, but for most model building it may only be necessary to use the auto-fitting facilities. You may need to fit the B conformation of a sidechain manually, since fitting with real-space refinement produces the same solution as the A conformation. This can be done using X-AUTOFIT | Build atoms | Refine 1 residue by gradient refinement, since this will move to the nearest density.
There are several ways to rebuild loops in X-AUTOFIT. If it is necessary to delete residues in a loop, use either Delete residue or Delete range on the X-AUTOFIT | Build atoms | Add delete palette to remove one or more residues. The new termini created by this process can be manually adjusted using X-AUTOFIT | Build atoms | Model first last 4 res. tool to bring them together or forced together using the X-AUTOFIT | Build atoms | regularize option. It will be necessary for the regularize option to work to insert a peptide bond between the two cut ends of the loop. Use the tool X-AUTOFIT | Build atoms | Add delete | Create peptide link to add this new bond; and since there is no length restriction on this new peptide bond, you can temporarily create a very long peptide bond, which will be regularized to the correct length.
To insert residues into a loop, you must first cut a peptide bond to create two new termini, using X-AUTOFIT | Build atoms | Add delete.
Delete peptide link. You can now use the X-AUTOFIT | Build atoms | Model first 4 res. tool to manually move the two ends to make room for the new residue(s) and then use the tool X-AUTOFIT | Build atoms | Add delete | Add res at termini to add the residues from either of the new cut ends. As well as deleting residues, a new peptide link can be created, and the new loop regularized to remove the strain. The new loop can be fitted as two cut ends before joining these together, refined with X-AUTOFIT | Structure | Refine zone, or searched with the tool X-AUTOFIT | Structure | Loop fit. Large changes can be searched with loop fitting algorithm, but there is a limit of around six residues on the size of loop that can be fitted in a reasonable period of time.
The Save changes and Undo last options under X-AUTOFIT | Build write and read a session file to and from disk. This is much quicker than restoring the data from an MSF file or writing a new file and allows you to save your editing in process and to undo a mistake very easily.
Rigid-body refinement (on the Structure palette) of a zone allows the fitting of sections of structure and very quickly finds the nearest minima for all atoms in the zone. The radius of convergence is not as great as in rigid-body refinement in reciprocal space, so it may not be able to improve rotation/translation function solutions for a whole protein. It may be possible to refine each domain or even secondary structure element.
The structure-building section includes rigid-body refinement and full torsion-angle refinement for a specified region of the structure. Thus it is possible to rigid-body refine a domain, segments, or even just local folds to improve the initial fit to a map. The refinement protocol uses torsion-angle refinement and rigid-body refinement on a residue basis while retaining geometry with constraints on bonds, angle, impropers, and nonbonds. This results in a refinement method with a very high radius of convergence (1.5 Å). This refinement protocol should not be used to replace normal xyz refinement, since it is only applied in real space and does not affect the map. It does provide a rapid method of refining regions using a very powerful algorithm. Loop fitting and terminal fitting are accomplished using a Monte Carlo sampling method that can screen thousands of conformations per second against the electron density while retaining the ten best current solutions during the search. The progress of the search is displayed, and you can interrupt it when the ten solutions have converged to a single solution or when an obvious fit can be observed.