2. Building and contributing maps¶
We are excited to collect pathway maps for every organism with a well characterized metabolic network. This section describes the process of building a new map, either from scratch or using a COBRA model with BiGG IDs.
2.1. Building from scratch¶
To build a map from scratch, you will first need a COBRA model for your map. See the section Escher, COBRA, and COBRApy for some background information on COBRA models.
If you would like to eventually contribute your map to the Escher website, it is important that your COBRA model adheres to the identifiers in the BiGG Database. Escher and BiGG are being developed together, and we want to maintain consistency and interoperability between them.
Once you have a COBRA model, you can follow these steps:
Load your model in the Escher Builder.
Begin building new reactions. If you are familiar with the genes in your organism, then try search for new reactions by their gene IDs.
Limit each map to ~200 reactions. Maps larger than this will slow down the Escher viewer, especially on old browsers. Rather than building one giant map, Escher is designed for building many, smaller subsystem maps.
When you have built a map for your a subsystem, save the map as JSON with a name that includes the model ID, followed by a period, followed by the name of the subsystem. For example:
iMM904.Amino acid biosynthesis.json
(Optional) Once you have a set of subsystem maps, you can set up a local Escher server so that subsystem maps appear in the “quick jump” menu in the bottom right corner of the screen (as seen here for iJO1366). To set this up, you will need to start a local server as describe in Running the local server. Next, find your local cache directory by running this command in a terminal:
python -c "import escher; print(escher.get_cache_dir(name='maps'))"
This will print the location of the local maps cache. Add your new subsystem maps to cache folder. Now, when you run the server (described in Running the local server), you should see that quick jump menu appear.
NOTE: The cache directory is organized into folders for organisms. You can use these folder for filtering by organism on the local launch page, or you can place the maps in the top directory.
NOTE 2: A similar approach can be used to access your models from the local launch page. Place maps in the folder indicated by:
python -c "import escher; print(escher.get_cache_dir(name='models'))"
2.2. Building from an existing map for a similar organism¶
Follow the instruction above, except, rather than starting from scratch, load an existing Escher map for a different organism.
Once you have the new model loaded, use the Update names and gene reaction rules using model button in the Model menu to convert all descriptive names and gene reaction rules in the model to those in the map. Reactions that do not match the model will be highlighted in red. (This can be turned off again in the settings menu by deselecting Highlight reactions not in model.)
Now, visit each highlighted reaction and see if you can replace it with an equivalent biochemical pathway from the model. If not, then delete the reaction and move on.
Finally, when there are no highlighted reactions left, you can repeat this for other subsystems.
2.3. Submitting maps to the Escher website¶
We have a repository for sharing maps, with few restrictions. Anyone can submit a pull request to add new maps, and you might find some of the existing maps there useful:
If you would like to contribute maps to Escher the main Escher website, this
takes a little more effort, but we appreciate it! You can make a Pull Request to
the GitHub repository escher.github.io. Make sure there is a folder with the
name of the organism in
1-0-0/maps. For example, a new yeast map goes in the
Then, name your map by concatenating the model ID and the map name, separated by a period. For example, a yeast map built with the genome-scale model iMM904 could be named:
iMM904.Amino acid biosynthesis.json
Then, add the JSON file for the model to the Pull Request if that model is not
already available. As before, make a folder for your organism within
1-0-0/models/. The model filename is just the model ID.
In this example, a correct Pull Request would include the following files:
1-0-0/maps/Saccharomyces cerevisiae/iMM904.Amino acid biosynthesis.json 1-0-0/models/Saccharomyces cerevisiae/iMM904.json