Predictive modelling of directed evolution for de-novo design of solid binding peptides

Abstract

Genetically Engineered Polypeptides for Inorganics are the solid binding polypeptides designed to exploit their molecular specificity and binding affinity towards certain inorganic material surfaces. These solid binding polypeptides are selected using combinatorial methods such as phage display. These selections need to be optimized using directed evolution. Directed evolution involves the application of the molecular insights gained from the previous methods to evolve the activities of extant peptides and proteins. In the current thesis, we have identified quantitative amino acid properties from the biopanning data for predicting directed evolution trends. We have also trained machine learning models for the modelling of the binding behaviours of 12 amino acid length MoS2 binding peptides, and for the de-novo design of sequences. Overall, we have developed a simple and efficient methodology for the predictive modelling of directed evolution for de-novo design of solid binding peptides. The protocols developed are expected to impact the technological applications on the peptide-single layer solid based bio/nano soft interfaces such as biosensors, bioelectronics, and potentially also medical applications.