(397c) PyRosetta: An Interactive Platform for Teaching Protein Structure Prediction and Design

Structures of proteins and protein complexes help explain biomolecular function, and computational methods provide an inexpensive way to predict unknown structures, manipulate behavior, or design new proteins or functions. However, current computational protein structure prediction and design packages are complex and difficult to learn. To make these approaches broadly accessible to biomolecular engineers with varied backgrounds, we have developed PyRosetta, a Python-based interactive platform for accessing the objects and algorithms within the Rosetta protein structure prediction suite. Rosetta, developed by a consortium of laboratories in the Rosetta Commons, has an unmatched variety of functionalities and is one of the most accurate protein structure prediction and design approaches (Das & Baker Ann Rev Biochem 2008; Gray Curr Op Struct Biol 2006). In PyRosetta, users can measure and manipulate protein conformations, calculate energies in low- and high-resolution representations, fold proteins from sequence, model variable regions of proteins (loops), dock proteins or small molecules, and design protein sequences. Furthermore, with access to the primary Rosetta optimization objects, users can build custom protocols for operations tailored to particular biomolecular applications. Since the Python-based program can be run within the visualization software PyMol, search algorithms can be viewed on-screen in real time. In this talk, we will detail how PyRosetta was used in an undergraduate elective course to teach both the fundamentals and the practical application of protein structure prediction and design.