Overview

This repo contains three high-level examples of Rosetta-PyData integration using the pyrosetta.distributed namespace:

• ddg_analysis - Development of a hybrid RosettaScripts-python modeling protocol, combining pyrosetta-based model preparation, RosettaScripts-based modeling and dynamic templating of RosettaScripts protocols.
• miniprotein_design - A hybrid batch-and-interactive workflow, demonstrating post-run analysis of a large-scale simulation.
• relax_benchmark - Baseline performance profiling of a simple dask implemented simulation via both multi-threaded and multi-node distributed clusters.

Non-Interactive Views

As these notebooks may include custom view components, particularly a 3dmol.js based viewer, we recommend using the nbviewer interface as a non-interactive way to view these components.

Interactive Environment

The analysis notebooks within this project rely on a conda environment specified in environment.yml. Conda packages for this environment are available on the Linux 64-bit platform.

Installation

1. Install and configure conda in your environment. We strongly recommend the miniconda installer.

2. Clone this git repository:

git clone https://github.com/proteininnovation/Rosetta-PyData_Integration

3. The pyrosetta package is available for linux and osx via the conda.graylab.jhu.edu conda repository. To access this channel, obtain a PyRosetta license. Then, update configure your .condarc to access the specified channel via:

conda config --add channels https://{USERNAME}:{PASSWORD}@conda.graylab.jhu.edu

4. Setup the working environment, including python, pyrosetta and jupyterlab via:

conda env create -f environment.yml -n rosetta_pydata_integration

Evaluation

Activate the working environment via:

conda activate rosetta_pydata_integration

Launch the jupyter interface via:

jupyter lab