What is it about?

Proteins in nature were generated following billions of years of evolution and therefore possess limited structural folds and biological functions. This study presents an open-source method to design new, well-folded protein structures through sequence-independent folding simulations. Although starting from native secondary structure assignments, the computationally created structures have novel tertiary folds significantly different from natural proteins in the PDB, highlighting the ability to explore novel areas of protein fold space.

Featured Image

Why is it important?

De novo protein design generally consists of two steps: structure and sequence design. Most current protein design studies have focused on sequence design with scaffolds adapted from native structures in the PDB, which renders novel areas of protein structure and function space unexplored. Our method enables automated design of new structural folds from any human-input secondary structure restraints and therefore enhances the potential to explore novel areas of protein structure and function through computational fold design.


Progress in the field of protein design has been encouraging, partly due to the increasing accuracy of sequence design simulations, especially with the recent introduction of AI and deep neural network learning. Many of the design studies, however, required structural scaffolds taken from existing structures in the PDB. We hope that our study can help solve the core problem of accurate fold design when starting from human-desired restraints, such as secondary structure assignments and distance/contact maps, shifting focus more toward the successful design of new protein folds and functions beyond natural proteins, which has been the long-term aim of the field.

Yang Zhang
National University of Singapore

Read the Original

This page is a summary of: De novo protein fold design through sequence-independent fragment assembly simulations, Proceedings of the National Academy of Sciences, January 2023, Proceedings of the National Academy of Sciences,
DOI: 10.1073/pnas.2208275120.
You can read the full text:




The following have contributed to this page