Protein sequence design with a learned potential.

Journal: Nature communications
Published Date: Feb 8, 2022

Abstract

The task of protein sequence design is central to nearly all rational protein engineering problems, and enormous effort has gone into the development of energy functions to guide design. Here, we investigate the capability of a deep neural network model to automate design of sequences onto protein backbones, having learned directly from crystal structure data and without any human-specified priors. The model generalizes to native topologies not seen during training, producing experimentally stable designs. We evaluate the generalizability of our method to a de novo TIM-barrel scaffold. The model produces novel sequences, and high-resolution crystal structures of two designs show excellent agreement with in silico models. Our findings demonstrate the tractability of an entirely learned method for protein sequence design.

Authors

  • Namrata Anand
    Department of Bioengineering, Stanford University, Stanford, CA, USA.
  • Raphael Eguchi
    Department of Biochemistry, Stanford University, Stanford, CA, USA.
  • Irimpan I Mathews
    Stanford Synchrotron Radiation Lightsource, Menlo Park, CA, 94025, USA.
  • Carla P Perez
    Biophysics Program, Stanford University, Stanford, CA, USA.
  • Alexander Derry
    Biomedical Informatics Training Program, Stanford University, Stanford, CA, USA.
  • Russ B Altman
    Departments of Medicine, Genetics and Bioengineering, Stanford University, Stanford, California, United States of America.
  • Po-Ssu Huang
    Department of Bioengineering , Stanford University , Shriram Center for Bioengineering and Chemical Engineering, 443 Via Ortega , Stanford , California 94305 , United States.

Keywords

External Resources

Popular Topics
Recent Journals