/trRosetta

A package to predict protein inter-residue geometries from sequence data

Primary LanguagePythonMIT LicenseMIT

trRosetta

This package is a part of trRosetta protein structure prediction protocol developed in: Improved protein structure prediction using predicted inter-residue orientations. It includes tools to predict protein inter-residue geometries from a multiple sequence alignment or a single sequence.

Contact: Ivan Anishchenko, aivan@uw.edu

Requirements

tensorflow (tested on versions 1.13 and 1.14)

Download

# download package
git clone https://github.com/gjoni/trRosetta
cd trRosetta

# download pre-trained network
wget https://files.ipd.uw.edu/pub/trRosetta/model2019_07.tar.bz2
tar xf model2019_07.tar.bz2

Usage

python ./network/predict.py -m ./model2019_07 example/T1001.a3m example/T1001.npz

Links

References

J Yang, I Anishchenko, H Park, Z Peng, S Ovchinnikov, D Baker. Improved protein structure prediction using predicted inter-residue orientations. (2020) PNAS. 117(3): 1496-1503