Tutorial for Gaussian Network Model Database (GNM-DB) – iGNM 2.0

 

1.  Query interface

2. Default Results page

2.1. Mode Shapes

2.2. Domain Separations by Dynamics

2.3. B-factors

2.4. GNM Connectivity Model

2.5. Cross-Correlations Map

2.6. Collectivity of slow modes

3. Results in plain text

 

1.  Query interface

    Enter the PDB ID (e.g., 101M) and click the “Go iGNM” button to submit a query to GNM database.

Both the GNM results for the biological assembly (BA) and the original PDB are provided in the current GNM database. JsMol is the default molecular visualizer for viewing the PDB structures color-coded by the size of motions in individual GNM modes or all modes. To view the structures with Jmol, instead, the up-to-date Java Runtime Environment is required. Please click the “Allow” button to run Jmol when the security warning displayed below pops up:

 

 

We recommend to check the box “Do not show this again for this app and website” before clicking the “Allow” button, to avoid seeing this window in future applications.

 

2. Default Results page

This page contains a J(s)Mol window (left), Results panel and the input information (right).

The J(s)Mol window displays the PDB structure colored based on the size of motions driven by the slowest two GNM modes (blue: almost rigid; and red: highly mobile, as indicated by the color code bar under the diagram).

Take the sperm whale myoglobin (PDB id: 101M) as an example. The corresponding Results page is displayed below.

 

 The Results panel contains six clickable items:

1) Mode shapes

2) Dynamics domain separations

3) B-factors (correlation between theoretical and experimental B-factors, for X-ray structures only)

4) GNM connectivity model

5) Cross-Correlations map

6) Collectivity of slow modes

 

The input information (Input file and parameters) contains input PDB ID and parameters used for GNM calculations. For example, the computations for PDB 101M were performed for all chains in the Biological Assembly, using the first model deposited in the PDB (PDB: 101M.pdb,    chain: all,    model No.: 1); 101M structure was resolved by X-ray crystallography, at a resolution of 2.7 Å (Exp. method: X-RAY DIFFRACTION (2.07 Å)); three atoms P, C4' and C2 were selected as network nodes representing nucleotides (No. of nodes for nucleotide: 3-node) . Note that the Cα atom was selected as the (single) node for each amino acid. Node pairs within cutoff distances of 7.3 and 7.3 Å  are adopted for amino acid pairs and for nucleotide pairs, respectively  (Cutoff for amino acid and nucleotide: 7.3 and 7.3 Å); the spring constant for contacting nodes is taken as unity, 1.

In addition, all the BA structures for the PDB are listed under the input information panel. Only the first BA of the PDB was pre-calculated. And, the others BA structures can be calculated using the online calculation engine by clicking the button of the list (it can be done in seconds for the structures with <1000 nodes).  For those PDBs which are as same as their first BA structures, the BA structures share the same GNM results with their original PDBs.

A corresponding button “GNM for the original PDB” is put in the same position to go back to the normal GNM result page. Take the PDB 3RW0 as an example:

 

2.1. Mode Shapes

Similar to the B-factors result page, 3D J(s)Mol windows and 2D interactive charts are provided to display the slow and fast modes along with the modes averages.

The 3D structure shown in J(s)mol is colored based on the mobility of the residues in a certain mode. The color spectrum goes from blue (most rigid) -> white -> red (most mobile).

 

The mode types (individual slow modes, average of the slowest 1-2 or 1-3, fast modes and fast average 1-10) can be selected from the drop-down menu. While the mode index can be selected only for the slow and fast modes. The default mode is the slow mode since the slow modes make the largest contribution to the fluctuation of the structure and usually correlate with biological function. The colors on the structure are updated automatically by changing the mode type or mode index. We provide two 3D windows for easy comparison between two modes.

 

For fast modes, the whole structure is colored blue except for the hot spot residues (those subjected to highest frequency fluctuations) which are colored red.

 

The modes in the 2D charts are scaled by the relative inverse of the eigenvalues of the Kirchhoff matrix (G). Similar to the B-factors result page, point labels and zoom in functionality are available for the interactive 2D chart. The Hide “All” button hides all the series in the chart making it easy to reset the chart.

 

 

The export buttons provide different ways to export customized figures or data in different formats. The activated buttons has an orange colored background while inactive buttons has a gray background.

 

2.2. Domain Separations

    The residues are separated based on the sign (+/-) of each residue in an eigenvector.  The residues with the same sign in the eigenvector moves together. The interfacial residues are shown using the wireframe representation. The 3D J(s)Mol windows and the 2D interactive chart have the same functionality as the Mode Shapes result page.

 

 

 

 

The labeled residues in the Domain Separation chart below are the interfacial residues whose neighboring residues have a different sign in the eigenvector relative to themselves. The interfacial residues act as hinges in the movement of the molecules.

 

 

2.3.  B-factors

The theoretical and experimental B-factors are displayed/compared using both 3D color-coded PDB structures using J(s)Mol and 2D interactive graphs. Corresponding PDB structure, color-coded by residue fluctuations (see below), can be downloaded from the “Download PDB” links.

In the 3D J(s)Mol window, the structures are represented as cartoon and color-coded by GNM-defined theoretical fluctuations or X-ray solved experimental B-factors (the B-factor column in PDB files). The colors are defined by the mobility of the residues/nodes. Rigid and mobile residues are colored blue and red respectively.

Click the “Export image” button to export images from the 3D J(s)Mol windows with different image types (png, jpg and gif) and resolution (600 px for both width and height by default). The pop-up export dialog box for JsMol is displayed as

while the export dialog box for JMol has more functions:

The image type, quality and compression can be selected with the dialog box.

 

Right click on the J(s)Mol windows to show the menu, and more functions can be found in the pop-up J(s)Mol menu:

 

The 2D profiles of B-factors as a function of residue index are plotted using the interactive chart. The x- and y-axis refer to the residue index (ID) and B-factors respectively.

The control panel below the chart can be used to hide/show the specified chain(s) and charts (theoretical and/or experimental). Also, the legends can be clicked to hide or show the corresponding curves in the chart. Three formats of images (png, jpg and svg) and PDF document of the customized 2D charts can be exported using the “Export” buttons. Also, all the text data in csv format can be export using the “CSV” export button.

 

Hover on the plots will display the plotted series with their names, the residue information (chain identifier, residue type and residue ID) and corresponding B-factors.

Besides, the points in the 2D graph can be clicked to label the corresponding residue in the 3D J(s)Mol window where Cα atoms are presented in gray balls and sidechain atoms are in “wireframe”.

 

 

 

 

 

Select a range of the 2D chart will zoom in the chart to selected view and a “Reset view” button will pop up in order to restore the view to molecule’s full length upon clicking.

 

 

2.4. GNM Connectivity model

The network of the GNM model is displayed in both 3D structure and 2D connectivity map. Each ball represents a node and each line between the nodes represents a spring connection/interaction between the pair of interest if the pair stays within a cutoff distance in space.

 

 

The color of atoms/nodes (balls) can be changed by clicking the radio button “Atom names” and “B-factors”. Similar to the 3D B-factors windows, images can be exported using the export button.

 

    The topology of the network can also be viewed from the 2D Connectivity map. Each dot represents a spring connection between residue i and j. The residue information of the residue i and j can be displayed while hovering the cursor on points in the map.

 

 

2.5. Cross-Correlations Map

The Cross-Correlation (CC) map for all modes is provided by default. For PDB structures with less than 1000 nodes, the CC map can be customized with any range of modes. A customized CC map can be obtained by changing the range of modes and then clicking the “Calculate” button.

Two CC maps (static and interactive, on the left and right hand side respectively) are shown for PDB structures with less than 1000 nodes. Hovering your mouse over the static (left) CC map, the information for the pair of residues of interest will be displayed below the map. By clicking on a point in the static map, the interactive CC map will show the zoomed in view containing 100 residues on both axis centered at the selected point in the static map. The detailed information can be displayed well in the interactive CC map for the big structures.

 

An example of interactive CC map of PDB ID: 101M in action when a point was selected in the static CC map:

 

The size of interactive CC map can be customized using the control panel below the map. The minimum residue indices i and j can be changed as well as the map size. The maximal indices are automatically calculated based on the minimal index and the map size (e.g., there are 100 points between min i 45 and max i 144). The “Redraw” button redraws the interactive CC map based on the settings (min i, j and the map size which define the map range) in the control panel. The “Refresh” button not only redraws the interactive CC map using the settings, but also reloads the page with a new URL address containing the new settings. For example, http://gnm.csb.pitt.edu/gnmdb/iGNM_CC_map.php?gnm_id=101M&modes=cc_all&x_min=4&x_max=103&axis_length=100&y_min=55&y_max=154&max_length=800 is the URL of the page generated by clicking the “Update” button. The new URL directly restores the customized view of the interactive CC map therefore it can be saved/shared for later viewing. The “Default” button resets the settings to the default settings.

 

Furthermore, the interactive CC map can be zoomed in by selecting a range of points on the map. The “Reset” button will reset the interactive CC map back to the original view.

 

 

 

    Selecting a point in the interactive CC map will label the Cα atoms (shown as spheres and same color as selected point) of the pair of residues in the 3D J(s)Mol window.   

 

2.6. Collectivity of slow modes

The degree of collectivity of each mode measures the degree of the segments of the structure that move together in that mode. The mode labelled with its value has the highest collectivity. The 2D interactive collectivity chart has similar functions as the B-factor 2D chart.  Hovering over the bar of each mode will display the mode number and its corresponding collectivity value.  Selecting a range of modes in the chart, a figure containing only the range of selected modes will be displayed. Lastly, the current figure and data can be exported in different formats using the export buttons.

 

3. Results in plain text

    The nodes files in PDB format, network topology and GNM Kirchhoff matrix used for the GNM calculation are listed. Also, the GNM outputs (eigenvalues, slow modes, fast modes, slow eigenvectors, fast eigenvectors, slow mode average 1-2, slow mode average 1-3 and fast mode average 1-10 are provided. Besides, the text format cross-correlations (for structures have nodes less than 1000 only), slow modes collectivity, theoretical and experiment B-factors along with the scaling factor between experimental B-factors and theoretical fluctuations and the estimated spring constant are provided in text format.

An example list for PDB 101M:

Nodes used for the calculation (.pdb)

Network topology (.txt)

GNM Kirchhoff matrix (Sparsed Form) (.kdat)

Cutoff distances based contact number of nodes (.cont)

Residues information (.ResInfo)

Number of the nodes (.txt)

Eigenvalues (.eigen)

Slow modes (.slowmodes)

Fast modes 1-20 (.fastmodes)

Slow eigenvectors (.sloweigenvectors)

Fast eigenvectors 1-20 (.fasteigenvectors)

Slow mode average 1-2 (.slow2av)

Slow mode average 1-3 (.slow3av)

Fast mode average 1-10 (.fast10av)

All modes Cross-Correlation (.txt)

Slow modes collectivity (.collectivity-slow)

Theoretical and experimental B-factor profiles (.bfactor)

Scaling factor between experimental B-factors and theoretical fluctuations (.txt)

Estimated value of spring constant (.txt)