Protein Optimization Evolving Tool (POET) based on Genetic Programming

Abstract

Proteins are used by scientists to serve a variety of purposes in clinical practice and laboratory research. To optimize proteins for greater function, a variety of techniques have been developed. For the development of reporter genes used in Magnetic Resonance Imaging (MRI) based on Chemical Exchange Saturation Transfer (CEST), these techniques have encountered a variety of challenges. Here we develop a mechanism of protein optimization using a computational approach known as “genetic programming”. We developed an algorithm called Protein Optimization Evolving Tool (POET). Starting from a small library of literature values, use of this tool allowed us to develop proteins which produce four times more MRI contrast than what was previously state-of-the-art. Next, we used POET to evolve peptides that produced CEST-MRI contrast at large chemical shifts where no other known peptides have previously demonstrated contrast. This demonstrated the ability of POET to evolve new functions in proteins. Interestingly, many of the peptides produced using POET were dramatically different with respect to their sequence and chemical environment than existing CEST producing peptides, and challenge prior understandings of how those peptides function. This suggests that unlike existing algorithms for protein engineering that rely on divergent evolution, POET relies on convergent evolution.

Development of a Synthetic Biosensor for Chemical Exchange MRI UtilizingIn SilicoOptimized Peptides

“…Consequently, this burden on metabolism reduces the reporter protein’s cellular concentration, leading to a decrease in overall contrast. To remedy this issue, we developed a machine learning algorithm that evolved peptides to optimize their CEST contrast while diversifying the amino acid residues in their sequences 40,43 . Having generated a wide range of 12 amino acid long peptides, questions about their expression in cells still existed.…”

“…Using the previously generated peptide library targeted for 3.6-ppm contrast from the Protein Optimization Engineering Tool 40 , an amino acid sequence was generated by connecting the top-performing peptides end to end to create a 198 (superCESTide 2.0) amino acid long protein. The amino acid sequence was then optimized for DNA expression in E. coli using the Azenta Life Sciences’ Codon Optimization tool on Genewiz.…”

“…To remedy this issue, we developed a machine learning algorithm that evolved peptides to optimize their CEST contrast while diversifying the amino acid residues in their sequences 40,43 .…”

The Proton Resonance Enhancement for CEST imaging and Shift Exchange (PRECISE) family of RF pulse shapes for Chemical Exchange Saturation Transfer MRI

“…In the present study, we have focused on developing versatile pulse shapes for CEST MRI using an applied a gradient ascent algorithm to optimize the shapes for a range of Δω and k ca that are observed in the agents employed in our center including the ones evaluated but also glucose 9 , superCESTide 37, 38 , creatine 39 and others. We focused on optimizing for a 100 msec saturation pulse to be inserted into a pulse train for producing CEST MRI contrast, so these pulses are not explicitly optimized for the steady-state CEST sequences with low flip angle pulse and buildup of saturation which are also widely used on clinical scanners.…”

Using Genetic Programming to Predict and Optimize Protein Function

“…However, since POET's exploration is not limited to specific parts of the search space, we managed to find proteins that do not follow this principle and yet have a high CEST contrast value. Details of this experiment can be found in [57]. Motifs with a single amino acid symbol are the most trivial to find and therefore are the most common motifs among the models.…”