Research Outline

Applications of AI for Directed Evolution in Biotechnology,

Goals

To case studies for the use of AI for directed evolution in biotechnology, with a focus on the AI products used. The research will be used to inform the process of building an AI and data analytics platform.

Early Findings

  • General applications of AI in Biotechnology include drug discovery & clinical trials, diagnostics, radio-therapy and radiology, personalized medicine, gene editing, electronic health record (EHR), and medication management
  • Directed evolution is the is an “iterative procedure which involves the identification of a starting state protein, diversification of its gene, an expression and screening strategy, re-diversification, re-screening, and so on until a satisfactory performance level in terms of enzymatic activity, binding affinity or specificity is reached.”
  • Machine learning comprises algorithms and statistical models that improve the computer performance of various tasks.
  • According to an article on Wiley Online Library, machine learning supports directed evolution of enzymes because it is a protein engineering process that generates huge amounts of potential training data.
  • Recent studies have demonstrated that the algorithms Innov'SAR and ASRA are suitable in guiding the performance of saturation mutagenesis at sites lining the binding pocket for enhancing stereoselectivity and activity.
  • Specifically, machine learning as a form of artificial intelligence perfectly helps when performing directed evolution of stereoselective enzymes based on focused saturation mutagenesis.
  • According to experts, “machine-learning approaches accelerate directed evolution by learning from the properties of characterized variants and using that information to select sequences that are likely to exhibit improved properties”.

Summary of Early Findings

  • The first hour of research yielded some useful insights into the applications of AI in directed evolution. However, we could not identify any pre-existing case study related to the same.
  • We believe that a more in-depth search may yield some case studies. If the same is not found, the research team will have to piece together information to create such case studies.
  • For the team to proceed with the research, please choose one or more of the proposed next steps in the scoping section below.