The open-source Masala software suite: Facilitating rapid methods development for synthetic heteropolymer design

Area of Science:

• Computational biology
• Biophysics
• Software engineering

Background:

• Protein design traditionally uses machine learning, but synthetic heteropolymer design relies on physics-based methods.
• The Rosetta software is a key tool but faces limitations due to its architecture and licensing.
• There is a need for modern, flexible, and open-source software for heteropolymer design.

Purpose of the Study:

• Introduce the Masala software suite as a successor to Rosetta for heteropolymer design.
• Provide a free, open-source, and modular C++ library to extend existing software.
• Accelerate protein and synthetic peptide design through efficient new modules.

Main Methods:

• Developed Masala as a modular C++ library with an automated build system for API layers.
• Implemented Masala modules as drop-in replacements for Rosetta's minimizer and packer.
• Re-implemented design-centric guidance terms for efficient feature promotion and discouragement.

Main Results:

• Masala provides a modular architecture enabling easy addition of new plugin modules.
• Masala modules offer significant speedups, up to two orders of magnitude, in benchmarks.
• Efficient re-implementation of guidance terms improves performance in heteropolymer design.

Conclusions:

• Masala offers a modern, flexible, and open-source alternative for advancing heteropolymer design.
• Its modularity and efficiency facilitate the development of new design methods.
• Masala is poised to become a valuable successor to Rosetta in the field.