Mechanism and Catalytic Site Atlas

The Output Page

The output page of each mechanism search can be accessed by clicking on the "Done" link in the overall EzMechanism table.

The output page contains four (optionally five) main panels. A graph of configurations, steps, and mechanistic paths is shown on the upper left (panel 1). The panel on the upper right (panel 2) contains buttons and sliders used to filter the graph and change its representation. Below the second panel, a third panel shows information about the catalytic rule if any reaction step (edge) is selected in the graph. Panel 4, located below the first three panels shows the 2D diagrams of selected configurations in the graph. Finally, panel 5 is only shown for searches made for enzymes that are already in the database, with the purpose of comparing the annotated mechanism with the results of the search.

In the following sections we explain how to interpret the results of an EzMechanism search and what are the capabilities of the output page.

1 - The Results Graph

In the results graph, configurations are represented as circles and reaction steps as the edges that connect two configurations. Each reaction step is associated with a single rule that generated that transformation.

The reactants and products configurations are coloured in orange and red, respectively, and labelled as "R" and "P". Configurations that were matched against the catalytic rules (explored configurations) are shown in yellow, and the remaining configurations, which were generated but not checked against the rules are colored in grey. The configurations are labelled with a unique identifier that starts with a number representing the distance to the reactants’ configuration (in terms of reaction steps), followed by a unique string of letters that hold no particular meaning.

Edges are typically colored in grey. However, if the prediction is for a mechanism already existing in the database, they are colored in red if the rule is exclusive to that database mechanism and colored in orange if the rule is seen in this mechanism but also elsewhere.

2 - Tools to Filter and Manipulate the Results Graph

The purpose of the buttons and slides on the second panel of the output page is to filter and trim the mechanism graph to facilitate comprehension. The three red buttons allow for the removal of any selected configuration, reaction step, or reaction rule (which removes all the reaction steps that follow the selected rule) from the graph. Any elements of the graph deleted in this manner are listed in the neighboring table and can be restored.

Below the three red buttons there is a slider to filter out any reaction steps that involve the formation of bonds between atoms farther away than the selected cut-off. Note that the distances are based on the position of the atoms in the PDB structure, so the formation of a bond between atoms that are 7 Å away for example, might not be unreasonable, since during the reaction the molecules in the active site might move to bring these atoms closer. A second slider allows the filtering-out of mechanistic paths that are longer that a chosen number of steps. By default, only the shortest paths are shown.

Below the sliders, there are four self-explanatory checkboxes to further limit the number of circles or edges in the graph. The second and third (“keep only rules from mechanism”, and “hide rules unique to mechanism”) are only relevant when the prediction is for a database entry with a curated mechanism.

Finally, the layout of the graph might be toggled between tree or network-like. The tree layout is particularly useful to see the length of reaction paths, and how far away each configuration is from the reactants. The network-like layout is useful to detect interesting shapes in the topology of the graph, such as bottleneck configurations or reaction steps that are essential to connect the reactants to the products.

3 - Rule Information

When an edge, representing a reaction step, is selected in the results graph, the rule that was used to generate that step is show in the third panel. Also shown, is the largest distance between atoms that form new bonds in the step, which is used to compute the final prioritization score, given inside parentheses. The iteration number shows how many configurations had to be explored before this reaction step was found. Finally, there is a link to a page that shows detailed information about the rule, including a listing of all the enzymes in the database that follow the rule.

4 - Schemes of the Configurations and Reaction Steps

The two-dimensional scheme of any configuration can be seen in panel 4 if the corresponding circle is selected in the graph. Similarly, if an edge is clicked, two configurations are shown, corresponding to the starting and ending points of that reaction step. In that case, the reaction centres (atoms that are matched to the rules and are involved in bond changes) of the step are highlighted in both schemes. The direction of the reaction step can be reversed for visualisation purposes by clicking a second time on the edge. When an edge is selected, all the edges of reaction steps that follow the same rule have their representation changed to a dashed line.

5 - Comparison with Annotated Mechanism

If these are the results for an EzMechanism search based on an existing entry of the database, the manually curated mechanism is shown in panel 5. The 2D curly arrow diagrams of every step of the mechanism are shown on the left side of the panel, so they can be compared with the output of the prediction. The catalytic rules that were originated from each step are shown on the right, with the indication if those rules were included in the prediction and matched any configuration during the search.