Skip to content
Permalink
Fetching contributors…
Cannot retrieve contributors at this time
424 lines (374 sloc) 14.9 KB
layout title questions objectives time_estimation key_points contributors
tutorial_hands_on
Visualizations: JavaScript plugins
How can I make a custom visualization plugin for Galaxy?
Learn how to add JavaScript plugins to Galaxy using the Charts visualization framework
1h
Demonstrating a pluggable extension system for JavaScript visualizations
With three primary files we can integrate any JavaScript visualization into Galaxy
shiltemann
yhoogstrate
bgruening
guerler
dannon

Introduction

{:.no_toc}

In this tutorial we are going to demonstrate how to add a third party JavaScript-based visualization to Galaxy, and we'll talk about what the benefits are. The plugin we've selected for this exercise is the PV-Javascript Protein Viewer. It is an open source protein structure viewer for PDB-files. There are many other popular protein structure viewers available for the visualization of PDB-files such as e.g. NGL (also available in Galaxy) and JSMol.

{% icon details %} Background: What is the PDB (Protein Data Bank) file format?

The PDB-file format contains atomic coordinates of biomolecules derived through a range of experimental and computational methods. Most commonly the file contains a spatial crystallographic snapshot of a protein. There are 100s of thousands of protein structures publicly available at the Protein Data Bank (https://www.rcsb.org). Proteins are usually labeled by a four-letter code. Here is an example of a PDB-file for a hydrolase bond to its inhibitor (PDB: 1ACB):

HEADER    HYDROLASE/HYDROLASE INHIBITOR           08-NOV-91   1ACB
TITLE     CRYSTAL AND MOLECULAR STRUCTURE OF THE BOVINE ALPHA-CHYMOTRYPSIN-EGLIN
TITLE    2 C COMPLEX AT 2.0 ANGSTROMS RESOLUTION
...
KEYWDS    SERINE PROTEASE, HYDROLASE-HYDROLASE INHIBITOR COMPLEX
AUTHOR    R.Z.KRAMER,L.VITAGLIANO,J.BELLA,R.BERISIO,L.MAZZARELLA,
AUTHOR   2 B.BRODSKY,A.ZAGARI,H.M.BERMAN
...
REMARK   1 REFERENCE 1
REMARK   1  AUTH   M.BOLOGNESI,L.PUGLIESE,G.GATTI,F.FRIGERIO,A.CODA,L.ANTOLINI,
REMARK   1  AUTH 2 H.P.SCHNEBLI,E.MENEGATTI,G.AMICONI,P.ASCENZI
REMARK   1  TITL   X-RAY CRYSTAL STRUCTURE OF THE BOVINE
REMARK   1  TITL 2 ALPHA-CHYMOTRYPSIN(SLASH)EGLIN C COMPLEX AT 2.6 ANGSTROMS
...
SEQRES   1 E  245  CYS GLY VAL PRO ALA ILE GLN PRO VAL LEU SER GLY LEU
SEQRES   2 E  245  SER ARG ILE VAL ASN GLY GLU GLU ALA VAL PRO GLY SER
SEQRES   3 E  245  TRP PRO TRP GLN VAL SER LEU GLN ASP LYS THR GLY PHE
...
ATOM      1  N   CYS E   1       2.323 -16.405  18.812  1.00 43.48           N
ATOM      2  CA  CYS E   1       3.017 -15.136  18.786  1.00 35.11           C
ATOM      3  C   CYS E   1       4.134 -15.068  19.799  1.00 32.90           C
ATOM      4  O   CYS E   1       4.173 -15.810  20.772  1.00 41.38           O
ATOM      5  CB  CYS E   1       2.052 -13.969  19.139  1.00 31.14           C
ATOM      6  SG  CYS E   1       1.246 -14.085  20.788  1.00 34.72           S
ATOM      7  N   GLY E   2       4.993 -14.081  19.607  1.00 21.94           N
...
HETATM 2292  O   HOH E 406      12.343   1.842  12.901  0.86 18.70           O
HETATM 2293  O   HOH E 407      -4.767  17.237  10.630  1.00 59.78           O
HETATM 2294  O   HOH E 408      11.489  -6.278  18.740  0.96 20.00           O
...

More resources on this file format:

As mentioned above we will be focusing on the PV-Javascript Protein Viewer in this tutorial. Now that we have learned about the underlying file format, let us continue by visiting the protein viewer developer site at https://biasmv.github.io/pv/ to get familiar with the plugin.

{% icon hands_on %} Hands-on

  1. View the plugin in action, rotate the molecule and change its style.

  2. Under which license is this plugin distributed?

  3. Can you find the minified code file of this plugin?

{: .hands_on}

Agenda

In this tutorial, we will deal with:

  1. TOC {:toc}

{: .agenda}

Section 1 - Basic plugin setup

Directory and plugin preparations

In this section we will download the viewer and add it to a local Galaxy instance. All development takes place within the Galaxy codebase. The first thing we are going to do is to clone a Galaxy instance and prepare the directory structure for the new visualization plugin.

{% icon hands_on %} Hands-on

  1. Clone an instance of Galaxy in a path, further referred to as $GALAXY_ROOT:

    $ git clone https://github.com/galaxyproject/galaxy
  2. Navigate to the visualization plugin directory:

    $ cd $GALAXY_ROOT/config/plugins/visualizations
  3. Copy the existing example into a new directory:

    $ cp -r example myviz

{: .hands_on}

This example visualization provides a great place to start with most of the basics already in place. Now that the directory structure is in place, let us review the example visualization. Each visualization contains a set of 3 files:

  • Logo (static/logo.png) which will appear in Chart's plugin selection interface.
  • Configuration (config/example.xml) describing input parameters and options.
  • Wrapper (src/script.js) which serves as a bridge between Galaxy and our 3rd-party plugin.

In the following sections we are going to discuss these files in more detail and modify them to incorporate a new visualization. Let's start with the logo for our visualization.

Your visualization needs a logo

Each visualization is represented by a logo in the Galaxy interface. This makes it easier for users to find and configure their visualization. The logo should be in the png-file format. It will appear with a width of 120 pixels.

Here's an example logo:

Logo

{% icon hands_on %} Hands-on

  1. Find an arbitrary image in PNG-file format. Possibly using Google's image search.

  2. Copy it to the myviz/static directory and name it logo.png. {: .hands_on}

Configure the visualization

Each visualization has a configuration file. In this case it is named example.xml. This file has conceptual similarities with a Tool's XML-file. It allows developers to specify a variety of attributes and input parameters for their visualization. Throughout this tutorial we are going to gradually augment this file but for now we keep it simple.

{% icon hands_on %} Hands-on

  1. Rename the file to config/myviz.xml

  2. Edit the file named config/myviz.xml and change the name and description.

  3. Go to the Galaxy root directory to install dependencies, activate the virtual environment, and invoke the Galaxy client build.

    $ cd $GALAXY_ROOT
    $ make setup-venv
    $ source .venv/bin/activate
    $ make client
  4. Run Galaxy

    $ ./run.sh

{: .hands_on}

Your visualization should now be loaded. You can verify that now by clicking on Visualize > Create Visualization in the top menu bar of Galaxy and finding your plugin with its new logo in the list there.

Assign a new datatype to your visualization

{% icon hands_on %} Hands-on

  1. Open the file named config/myviz.xml and find the <data_source> section.

  2. Replace the existing two <test> sections with:

    <test type="isinstance" test_attr="datatype" result_type="datatype">molecules.PDB</test>

  3. Remove the <specs> and <groups> sections.

{: .hands_on}

This links the plugins to the PDB-file format, which means that for any history item of this file type the plugin will automatically be available.

Modifying the wrapper

Now let's take a look at the wrapper which connects our visualization with Galaxy. The wrapper consists of a module written in JavaScript and is available at src/script.js:

The wrapper receives an options dictionary with the following four components:

  • charts: The model of the current visualization with attributes, settings etc.
  • process: A jQuery.Deferred() object to allow asynchronous data requests within the wrapper
  • dataset: Details on the selected datasets such as url, ids etc. which can be used to access the dataset
  • targets: The DOM ids of the container elements to draw into

In this tutorial we will implement the PV-Viewer plugin. In order to execute a 3rd-party plugin we need to figure out how it works. This can be done by finding a working example or documentation. Fortunately the PV-Viewer comes with both. Let's take a look at the documentation.

{% icon hands_on %} Exploring the PV-Viewer

  1. Identify the parameter which is needed to initialize the plugin when calling pv.Viewer().

  2. Which of the wrapper option components represents this parameter?

  3. Can you identify which mode settings are valid to render the structure with pv.Viewer.renderAs()?

{: .hands_on}

Now that we have learned the basics on how the viewer plugin works, we can edit it and adjust script.js.

{% icon hands_on %} Hands-on

  1. Access your visualization's myviz/src directory.

    $ cd $GALAXY_ROOT/config/plugins/visualizations/myviz
  2. Install the package for the PV-Viewer:

    yarn add bio-pv
  3. Modify your plugin's entrypoint script src/script.js to import the new bio-pv library we just added, add this at the top of the file:

    import * as pv from 'bio-pv';
  4. Lastly, replace the load function contents in the same entrypoint file with the following code that will set up the protein viewer:

    var viewer = pv.Viewer( document.getElementById( options.targets[ 0 ] ), {
        width       : 'auto',
        height      : 'auto',
        antialias   : true,
        outline     : true
    });
    $.ajax( {
        url     : options.dataset.download_url,
        success : function( response ) {
            var structure = pv.io.pdb( response );
            viewer.clear();
            viewer.renderAs( 'protein', structure, 'cartoon', {} );
            viewer.centerOn( structure );
            viewer.autoZoom();
            options.chart.state('ok', 'Chart drawn.');
            options.process.resolve();
        }
    });

{: .hands_on}

Build the package

Now that we have completed the basic plugin definition, it is time to build the scripts and libraries into a single bundle that Galaxy can use. Galaxy visualizations typically use Parcel for this due to its simplicity, and this is how the example project is already configured. If you're interested in more details, take a look at the package.json file in your myviz plugin directory.

After it has been built and staged the plugin will be accessible through Galaxy's user interface. This process not require restarting your Galaxy instance, just make sure to properly refresh your browser.

{% icon hands_on %} Hands-on

  1. Navigate to your visualization's root directory:

    $ cd $GALAXY_ROOT/config/plugins/visualizations/myviz
  2. Install any necessary JavaScript dependencies:

    $ yarn install
  3. Run the plugin build process using yarn:

    $ yarn run build
  4. Stage the scripts and run Galaxy:

    $ cd $GALAXY_ROOT
    $ make client
    $ ./run.sh

{: .hands_on}

Lets test this.

Test the visualization

In this section we will select a PDB-file from the Protein Data Bank and visualize it with our new plugin.

{% icon hands_on %} Hands-on

  1. Visit https://www.rcsb.org and select a protein structure e.g. 1ACB

  2. Copy the link to the raw PDB-file e.g.

    https://files.rcsb.org/view/1ACB.pdb
  3. Start your Galaxy instance

    $ cd $GALAXY_ROOT
    $ ./run.sh
  4. Open the upload dialog, paste the above link and click on Start.

  5. Close the upload dialog, and select the file from the history panel on the right.

  6. Click on the diagram icon. (You must be logged in)

  7. Find your visualization and click on its logo.

{: .hands_on}

First view

Section 2 - Allow different rendering modes

In this section we are going to augment the visualization such that users may select different rendering modes. Similar to a Tool's XML file, developers may specify input parameters which then will be presented to the user. The definition of Tool and Visualization input parameters are similar, however the latter is provided in JavaScript and not as XML.

More information on parameters can be found in the wiki.

{% icon hands_on %} Hands-on

  1. Add the following block into the myviz.xml file:

    <settings>
       <input>
           <name>mode</name>
           <label>Display mode</label>
           <type>select</type>
           <display>radio</display>
           <value>cartoon</value>
           <help>Select the rendering mode.</help>
           <data>
               <data>
                   <value>cartoon</value>
                   <label>Cartoon</label>
               </data>
               <data>
                   <value>lines</value>
                   <label>Lines</label>
               </data>
               <data>
                    <value>points</value>
                    <label>Points</label>
               </data>
           </data>
       </input>
    </settings>
  2. Change the following line in script.js:

    viewer.renderAs( 'protein', structure, 'cartoon', {} );

    to

    var settings = options.chart.settings;
    viewer.renderAs( 'protein', structure, settings.get( 'mode' ), {} );
  3. Rebuild the plugin

    $ yarn run build
    $ cd $GALAXY_ROOT
    $ make client
  4. Refresh your browser.

  5. Load your visualization and test different rendering modes in the Customization tab of your visualization.

{: .hands_on}

Conclusion

{:.no_toc}

First of all, thank you for completing this tutorial. We have learned how to add JavaScript visualizations to Galaxy utilizing the Charts framework.

You can’t perform that action at this time.