Prot2Prot

Prot2Prot is a deep-learning model that imitates a Blender-rendered molecular image given a much simpler representation ("sketch") of a protein surface. Prot2Prot outputs an image that is often indistinguishable from a BlendMol-based visualization in a fraction of the time, allowing image "rendering" even in a web browser.

Prot2Prot is released under the terms of the Apache License, Version 2.0. See LICENSE.md for details.

Usage

Prot2Prot Web App

The Prot2Prot web app works on all major operating systems. Most users should simply visit the Prot2Prot website to use the model in their browser.

You can also download a copy of the Prot2Prot web app (including the Prot2Prot models themselves) to run on your own server.

Command-line-interface Prot2Prot

Comand-line-interface (CLI) Prot2Prot requires Linux. Users can compile and run CLI Prot2Prot by following these steps:

# Clone the repository
git clone https://git.durrantlab.pitt.edu/jdurrant/prot2prot.git

# Change into the prot2prot directory
cd prot2prot/

# Install the third-party dependencies
npm install

# Change into the nodejs directory
cd nodejs/

# Compile command-line prot2prot
./compile.sh

# Change into the newly created dist/ directory
cd dist/

# You must separately download and uncompress the trained prot2prot models
wget http://durrantlab.com/prot2prot/prot2prot_models.zip
unzip prot2prot_models.zip
rm prot2prot_models.zip

# You can now use the CLI using the render_*.js files
ls render_*js

# For help, use the -h flag like this:
node render_still.js -h

# See ./tests.sh for examples of use
./tests.sh