next up previous contents
Next: Interface and controls Up: Timeline Tutorial Previous: Contents   Contents



This tutorial demonstrates how to use the VMD plugin Timeline to analyze and identify events in molecular dynamics (MD) trajectories. Timeline creates an interactive 2D box-plot - time vs. structural component - that can show detailed structural events of an entire system over an entire MD trajectory. Events in the trajectory appear as patterns in the 2D plot. The plugin provides several built-in analysis methods, and the means to define new analysis methods. Timeline can read and write data sets, allowing external analysis and plotting with other software packages. Timeline includes features to help analysis of long trajectories and trajectories with large structures.

In the main 2D box-plot graph, users identify events by looking for patterns of changing values of the analyzed parameter. The user can visually identify regions of interest - rapidly changing structure values, clusters of broken bonds, differences between stable and non-stable values, and similar. The user can explore the resulting structures by tracing the mouse cursor (``scrubbing'') over the identified areas. The structure is highlighted and the trajectory is moved in time to track the highlight.

If you have any questions or comments on this tutorial, please email the TCB Tutorial mailing list at The mailing list is archived at

Getting Started

If you downloaded the tutorial from the web, the files that you will be needing can be found in a directory called timeline-tutorial-files. If you received the files through a workshop, they can be found in the same directory under the path $\sim$/Workshop/timeline-tutorial.

This will place you in the directory containing all the necessary files. In the figure below, you can see the structure of this directory.

Figure 1: Directory structure for tutorial exercises. Output for file-writing exercises is provided in the ``example_output'' subdirectory.

Figure 2: Exploring a secondary structure event. Row a) shows the system before the local structure change, row b) shows the system afterwards. The left panels show the entire data plot. The middle panels shows a zoomed-in part of the data plot, and the right panels shows the 3D structure of the selected trajectory frame. The vertical bar indicates the current frame. The purple oval, added for this figure, shows the rough area of the identified event. See text for additional details.

Required Programs

The following programs are required for this tutorial:

Timeline in action: examine events in titin domain extension

This subsection provides a short run-through of of Timeline features to provide a general idea of how the plugin is used. Later sections will introduce the display and user interface in greater detail. Here, we look at an MD trajectory of the I91 domain of the muscle protein titin. In the simulation trajectory, an external force is applied to the I91 domain with the Steered Molecular Dynamics method: one terminus of the domain is fixed and a harmonic restraint moving with constant velocity is applied to the other terminus, causing the domain to extend, unfold, and unravel. The secondary structure changes greatly as the domain unfolds; examining these changes will give us some initial insight into how the domain architecture responds to applied force.

With the titin I91 extension trajectory loaded into VMD, the user starts Timeline by selecting Extension $\rightarrow$ Analysis $\rightarrow$ Timeline from the main VMD window, then generates the secondary structure plot by selecting Calculate $\rightarrow$ Calc. Sec. Structure in the VMD Timeline window.

We examine trajectory events which appear as changes in value in the 2D plot. As the user moves the highlight cursor around the area of an event in the 2D plot, she can explore the behavior of the 3D structure. Figure 2 illustrates examining an event in a Timeline plot of per-residue secondary structure. By moving the cursor around the event area indicated with a purple oval, the user observes a beta strand, depicted in yellow, losing its secondary structure as it peels away from another beta strand. Note that the purple oval is added for the purposes of this tutorial, the user must normally visually identify the event area. Here the event to notice is the abrupt transition of a horizontal region of yellow to white: part of the domain losing the beta strand character (yellow) it had since the start of the trajectory and becoming random coil (white) as the domain extension proceeds. A user would scrub the mouse highlight cursor around the area of the purple oval, would produce the state shown in Figure 2a, before the local structure change, and examine the steps needed to reach Figure 2b, after the local structure change has taken place. To ``scrub'', click down the left mouse button with the cursor in the area of interest in the 2D graph, then move the cursor around the area of interest with the button still depressed - while splitting attention between the 2D graph, the changing 3D structure, and the changing numerical values displayed in the Highlight Details panel (described below).

Figure 3: Exploring a salt bridge breaking event. A Timeline plot of the salt bridges found throughout a trajectory. The structural elements in the vertical axis are selection groups, with each group being the pair of residues involved in a salt bridge. The left panels show the data plot, and the right panels show the 3D structure of the selected trajectory frame. The vertical bar indicates the current frame. The purple oval, added to this figure, shows the rough area of the identified event. Row a) shows the system before the salt bridge breaks, row b) shows the system after the salt bridge breaks. See text for additional details.

In Figure 3, a plot of salt bridge lifetimes during the titin domain extension trajectory shown in Figure 2 is displayed, illustrating moving the highlight cursor to examine a salt bridge breaking event. This plot is per-selection; each selection is the pair of residues involved in a salt bridge. Only residue pairs that form a salt bridge for at least one frame of the trajectory are listed. A highlighted salt bridge is seen breaking between Figure 3a and 3b. The Timeline highlighted representation can be manipulated like any other VMD representation; here coloring of the highlight was changed from the default ColorID 1/red to ResType to show the different residue types of the salt bridge partners, and easily track salt bridge partners as they move apart.

Figure 4: Using the threshold count to find a color scaling range. The per-residue Root Mean Square Fluctuation (RMSF) plot in a) has color scale set to the default, i.e., full, data range. Adjusting the Threshold Count range sliders in b) shows multiple peaks in the for the displayed range in c). In d), the 2D plot coloring scale of has been set to the range that was found in c), revealing which structures contribute to the threshold count.

In Figure 4 Timeline's Threshold Count plot is used to help find a suitable color scaling for the plot shown (per-residue Room Mean Square Fluctuation (RMSF)). The Threshold Count is the number of residues or selections in each frame with a value within a given range. As the Threshold Count range controls are adjusted as shown in Figure 4b, a pattern with multiple peaks appears (Figure 4c), this range is then used to change the color scale, with the result shown in Figure 4d. The Threshold Count sliders allow the users to very quickly scan through threshold ranges, without taking the time to redraw the entire 2D plot, while still giving an overview of the entire trajectory.

Per-residue vs. per-selection calculations

Timeline plots data about a list of structural elements vs. time. There are two ways of defining the list of structural elements to be analyzed in Timeline.

Per-residue: a list of residues; the protein and nucleotide residues in the user-defined selection. The default selection is ``all'', so this defaults to being a list of all residues in the current molecule.

Per-selection: a list of any VMD selections. For example, a list of salt bridge pairs, a list of the segments in a large protein complex, a list of of favorite structures in a molecule. These may be defined in built-in functions, listed explicitly, or defined algorithmically in user scripts.

next up previous contents
Next: Interface and controls Up: Timeline Tutorial Previous: Contents   Contents