Media, methods, and systems for protein design and optimization

Abstract

Exemplary embodiments relate to a protein engineering pipeline configured to optimize or improve proteins for specified functions. The problem space of such a task can grow quickly based on the sequence of the protein being optimized and the functions for which the protein is being designed. The solutions described herein allow the problem space to be efficiently searched by applying a combination of a protein design pipeline and an evaluation procedure performed on a quantum computer. As a result, single or multiple amino acid substitutions at a site of interest may be predicted in order to generate optimized protein variants.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 providing, to a processor, a protein sequence for optimization;   providing, to the processor, a protein structure having the protein sequence;   providing, to the processor, a protein property for optimization;   defining a scoring function based on the protein property for optimization;   providing, to the processor, at least one of:
 a position in the protein sequence to be subjected to modification, or an amino acid to be substituted for the amino acid occurring at the position, inserted at the position, or deleted from the position, or 
 a rotamer library to be searched for a target rotamer to be applied to one or more positions in the protein sequence; 
   determining, by the processor, a search space based on the at least one of the position, amino acid substitution, amino acid insertion, amino acid deletion, or rotamer from the rotamer library;   searching the search space using a quantum computing algorithm based on the scoring function, the searching comprising identifying at least one of a point mutation to the protein sequence or a combination of mutations to the protein sequence; and   providing an output state of the quantum computing algorithm, the output state indicative of an optimized protein sequence of an optimized protein, the optimized protein sequence being optimized according to the scoring function based on the protein property for optimization.   
     
     
         2 . The method of  claim 1 , further comprising:
 providing, to the processor, a predefined binding partner of interest, and wherein providing the protein having the protein sequence for optimization comprises applying a computer-based model based on a structure or amino acid sequence of the binding partner of interest, or   optimizing a binding partner for the selected protein during the searching.   
     
     
         3 . The method of  claim 1 , further comprising:
 testing the optimized protein for the protein property for optimization;   generating experimental data from the testing; and   applying an artificial intelligence method to the experimental data to learn an association between a configuration of the optimized protein sequence and the protein property for optimization, the configuration comprising a two- or three-dimensional protein structure, an amino acid sequence, a DNA sequence that encodes the protein, or parts of a two- or three-dimensional protein structure, in particular a catalytic domain, or one or more domains of a protein.   
     
     
         4 . The method of  claim 3 , further comprising providing, based on the learned association, a protein sequence, a position, an amino acid substitution, an amino acid deletion, an amino acid insertion, or a rotamer library for consideration for the search space. 
     
     
         5 . The method of  claim 1 , wherein the quantum computing algorithm is a quantum annealing algorithm configured to search the search space for the optimized protein based on a target Hamiltonian determined from the scoring function. 
     
     
         6 . The method of  claim 1 , wherein the quantum computing algorithm is one of a quantum-inspired algorithm, digital annealing algorithm, quantum annealing algorithm, gate-based quantum algorithm, quantum simulation algorithm, or a quantum-inspired optimization. 
     
     
         7 . The method of  claim 1 , further comprising determining the protein sequence for optimization based on the protein property for optimization via an artificial intelligence or machine learning algorithm. 
     
     
         8 . The method of  claim 1 , wherein the protein sequence for optimization includes an amino acid or a DNA sequence. 
     
     
         9 . The method of  claim 1 , further comprising providing, to the processor, multiple target rotamers to be applied to a plurality of positions in the protein sequence. 
     
     
         10 . The method of  claim 1 , further comprising providing, to the processor, a plurality of positions in the protein sequence to be subjected to modification, or a plurality of amino acids to be substituted for the amino acids occurring at the positions, inserted at the positions, or deleted from the positions. 
     
     
         11 . The method of  claim 1 , wherein the output state of the quantum computing algorithm is indicative of a plurality of optimized proteins, and further comprising:
 ranking, by the processor, each optimized protein of the plurality of optimized proteins;   determining, by the processor, a subset of the ranked plurality of optimized proteins; and   further performing an optimization of a protein of the subset of the ranked plurality of optimized proteins.   
     
     
         12 . A system comprising:
 a classical computing system comprising a non-transitory computer-readable medium and a processor configured to:
 determine a protein having a protein sequence for optimization; 
 receive a protein property for optimization; 
 define a scoring function based on the protein property for optimization; 
 determine at least one of:
 a position in the protein sequence to be subjected to modification, 
 a replacement amino acid to be substituted for, an amino acid occurring at the position, 
 an amino acid deletion at the position, 
 an amino acid insertion at the position, or 
 a rotamer library to be searched for a target rotamer to be applied to the protein sequence; and 
 
 determine a search space based on the at least one of the position, amino acid substitution, amino acid insertion, amino acid deletion, or the target rotamer from the rotamer library to be applied to the protein sequence; and 
   a quantum computing system comprising a plurality of qubits, a qubit control device, and a measurement unit configured to search the search space using a quantum computing algorithm based on the scoring function, and to provide an output of the quantum computing algorithm, the output indicative of an optimized protein, the optimized protein being optimized according to the scoring function based on the protein property for optimization.   
     
     
         13 . The system of  claim 12 , wherein the system is further configured to:
 receive a predefined binding partner of interest, and wherein to determine a protein having a protein sequence for optimization the classical computing system is further configured to apply a computer-based model based on a structure or amino acid sequence of the binding partner of interest, or   optimize a binding partner for the selected protein during the searching.   
     
     
         14 . The system of  claim 12 , wherein the system is further configured to:
 receive experiment data from testing the optimized protein for the protein property for optimization; and   apply an artificial intelligence method to the experimental data to learn an association between a configuration of the optimized protein sequence and the protein property for optimization, the configuration comprising a two- or three-dimensional protein structure, an amino acid sequence, a DNA sequence that encodes the protein, or parts of a two- or three-dimensional protein structure, in particular a catalytic domain, or one or more domains of a protein.   
     
     
         15 . The system of  claim 14 , wherein the system is further configured to provide, based on the learned association, a protein sequence, a position, an amino acid substitution, an amino acid deletion, an amino acid insertion, or a rotamer library for consideration for the search space. 
     
     
         16 . The system of  claim 12 , wherein the quantum computing algorithm is a quantum annealing algorithm configured to search the search space for the optimized protein based on a target Hamiltonian determined from the scoring function. 
     
     
         17 . The system of  claim 12 , wherein the quantum computing algorithm is one of a quantum-inspired algorithm, digital annealing algorithm, quantum annealing algorithm, gate-based quantum algorithm, quantum simulation algorithm, or a quantum-inspired optimization. 
     
     
         8 . The system of  claim 12 , wherein the protein sequence for optimization includes an amino acid or a DNA sequence. 
     
     
         19 . The system of  claim 12 , wherein the output of the quantum computing algorithm is indicative of a plurality of optimized proteins, and wherein the system is further configured to:
 rank, by the processor, each optimized protein of the plurality of optimized proteins;   determine, by the processor, a subset of the ranked plurality of optimized proteins; and   further perform an optimization of a protein of the subset of the ranked plurality of optimized proteins.   
     
     
         20 . A non-transitory computer-readable medium storing instructions, that when executed, cause a system to:
 determine a protein having a protein sequence for optimization;   receive a protein property for optimization;   define a scoring function based on the protein property for optimization;   determine at least one of:
 a position in the protein sequence to be subjected to modification, 
 a replacement amino acid to be substituted for, an amino acid occurring at the position, 
 an amino acid deletion at the position, 
 an amino acid insertion at the position, or 
 a rotamer library to be searched for a target rotamer to be applied to the protein sequence; and 
   determine a search space based on the at least one of the position, amino acid substitution, amino acid insertion, amino acid deletion, or the target rotamer from the rotamer library to be applied to the protein sequence;   search the search space using a quantum computing algorithm based on the scoring function, the searching comprising identifying at least one of a point mutation to the protein sequence or a combination of mutations to the protein sequence; and   provide an output state of the quantum computing algorithm, the output state indicative of an optimized protein sequence of an optimized protein, the optimized protein sequence being optimized according to the scoring function based on the protein property for optimization.