Machine Learning Models Revolutionize Targeted Cell and Gene Therapies

Strategic Insights -

Researchers at Stanford University are leveraging machine learning to improve the efficacy and safety of targeted cell and gene therapies. By using machine learning algorithms to predict which proteins are more likely to trigger an immune response, the team has made significant progress in designing proteins that avoid such issues and maintain their functionality when introduced into the human body. The breakthrough involves the use of zinc fingers, tiny proteins that regulate gene expression, and has the potential to revolutionize the field of cancer gene therapy.

Key Takeaways:

  • The Gao Lab at Stanford University has developed a machine learning-based approach to designing proteins that avoid immune responses, with a focus on zinc fingers, tiny proteins responsible for regulating gene expression.
  • The team combined three independent machine learning algorithms, including MARIA, a model designed to predict the immunogenicity of zinc protein junctions, and ESM-IF1, a protein language model that suggests targeted mutations to improve zinc finger performance.
  • The approach involves predicting new DNA targets that bind to combinations of zinc fingers, assembling them into arrays to create new junctions, and then leveraging MARIA to screen for junctions that avoid immune detection.
  • The team's results showed that the AI-enhanced zinc fingers increased gene expression by 2- to 6-fold, compared to the original proteins.
  • The researchers plan to build on this method, aiming for an end-to-end algorithm that could someday help design zinc-finger gene therapies for humans.
  • The study was funded by the National Institutes of Health, Longevity Impetus Grants, Stanford ChEM-H Seed Grant, and the Stanford Bio-X Interdisciplinary Initiatives Seed Grant Program.

Statistics:

  • The team's approach has made significant progress in designing proteins that avoid immune responses, with a focus on zinc fingers.
  • The combined use of three machine learning algorithms, MARIA and ESM-IF1, has consistently improved the functionality and reduced the immunogenicity of zinc fingers.
  • The AI-enhanced zinc fingers increased gene expression by 2- to 6-fold, compared to the original proteins.
  • The study aims to develop an end-to-end algorithm that could someday help design zinc-finger gene therapies for humans.

Sources:

  • "Machine learning-based approach to designing proteins that avoid immune responses" published June 3 in Cell Systems
  • National Institutes of Health, Longevity Impetus Grants, Stanford ChEM-H Seed Grant, and the Stanford Bio-X Interdisciplinary Initiatives Seed Grant Program
  • Stanford University, Gao Lab, School of Engineering, School of Medicine
  • "Why not design treatments that avoid immune reactions from the start?" - Xiaojing Gao, senior author and assistant professor of chemical engineering in the School of Engineering at Stanford
  • "We have taken the engineering of zinc fingers to a hitherto unvisited place, while simultaneously conserving function and lowering immunogenicity" - Xiaojing Gao, senior author and assistant professor of chemical engineering in the School of Engineering at Stanford