Evolution Of Models To Predict Protein Function Using Genetic Programming




Genetic Programming (GP) follows the concept put forth by Darwin – “Survival of the Fittest”. Species that evolve and adapt in response to the environment are more likely to survive. GP evolves models that can optimize peptides and proteins for variety of applications. It requires only basic understanding in protein engineering. The output can be a peptide or a protein that has improved function and yet, has a sequence that is significantly different from the original sequence. One application is to optimize peptides that can be visualized by magnetic resonance imaging (MRI). Specifically, the chemical exchange saturation transfer (CEST) contrast produced by a given protein’s amino acid sequence can be optimized through the use of variables, operators and functions to solve and optimize a problem. Sequence pattern tables are used to represent the GP model. The sequence pattern tables are weighted in importance of that pattern in order to achieve a better CEST contrast. For this technology, CEST is used to determine efficacy of the protein. Depending on the protein of interest, other methods to test efficacy will be used. Translation tables are used to translate amino acid letters into a value which describes a property of that amino acid which allows the GP to explore different search dimensions.




This invention is the development of the Protein Optimization Engineering Tool (POET) which optimizes protein function using Genetic Programming. POET evolves models that can predict protein functions from their sequence and can optimize proteins by predicting potential peptides with regard to special functions. These peptides can be tested experimentally to determine accuracy and added to the training dataset to evolve more accurate models for future experiments.



  • Simplified algorithm for optimizing protein function



  • MRI imaging
  • Drug Delivery
  • Drug and antibiotic discovery
  • Biomaterial development







Patent Information:

For Information, Contact:

Anupam Jhingran
Technology Manager
Michigan State University