Awesome
<img src="EVzoom_logo.png" alt="EVzoom logo" width="400"/>Take it for a spin @ https://EVzoom.org
EVzoom is an interactive, embeddable tool for visualizing undirected graphical models of protein families. Since these models explicitly parameterize all possible combinations of amino acids at all pairs of positions in a sequence, they contain far too much information to depict statically.
EVzoom lets you zoom in on any pair of positions in a protein family and see both inferred couplings between amino acids and sequence logos that summarize conservation statistics.
<p align="center"><img src="https://marks.hms.harvard.edu/evzoom/evzoom.gif" width="500"></p>EVzoom is based on SVG and powered by D3.
Inference
EVzoom is designed to be used with models inferred by plmc. The examples in the plmc repo show how to export JSON-formatted model files for EVzoom.
Want to work directly with the couplings in an EVzoom visualization? Check out the EVmutation Python package written by Thomas Hopf.
Embedding
Embedding EVzoom takes two lines
<div id="evzoom-viewer" data-couplings="/data/dhfr.json"> </div>
<script src="dist/evzoom.js"></script>
The data-couplings tag specifies the URL for the json file. This tag can be overridden by appending a query string ?data=JSON_URL to the URL. An example of the tag-based approach is available in example/evzoom.html. To take it for a spin and serve it from your filesystem, run python -m SimpleHTTPServer 8000 in the root of the repository (requires Python 2.7) and navigate to localhost:8000/example/evzoom.html in your browser.
Author
EVzoom was written by John Ingraham in Debora Marks' lab at Harvard Medical School