Machine learning method for protein modelling to design engineered peptides

Abstract

Provided herein are methods for design of engineered polypeptides that recapitulate molecular structure features of a predetermined portion of a reference protein structure, e.g., an antibody epitope or a protein binding site. A Machine Learning (ML) model is trained by labeling blueprint records generated from a reference target structure with scores calculated based on computational protein modeling of polypeptide structures generated by the blueprint records. The method may include training an ML model based on a first set of blueprint records, or representations thereof, and a first set of scores, each blueprint record from the first set of blueprint records associated with each score from the first set of scores. After the training, the machine learning model may be executed to generate a second set of blueprint records. A set of engineered polypeptides are then generated based on the second set of blueprint records.

What is claimed is: 
     
         1 . A method for designing engineered polypeptides using a machine learning model, comprising:
 receiving a representation of a reference target structure for a reference target;   generating a training set of blueprint records from a predetermined portion of the reference target structure, wherein each blueprint record comprises one or more target residue positions and one or more scaffold residue positions, each target residue position corresponding to one target residue from the one or more target residues;   labeling each blueprint record of the training set of blueprint records with a score;   training a machine learning model based on the training set; and   applying the trained machine learning model to a set of desired scores to generate an output set of blueprint records with one or more desired scores, wherein the output set of blueprint records are used to design the engineered polypeptides.   
     
     
         2 . The method of  claim 1 , wherein in at least one blueprint record, the residue portions are nonconsecutive. 
     
     
         3 . The method of  claim 1 , further comprising generating one or more polypeptides sequences from a blueprint record of the output set of blueprint records by computational protein modeling. 
     
     
         4 . The method of  claim 1 , further comprising generating one or more engineered polypeptides having a polypeptide sequences generated from the output set of blueprint records by computational protein modeling. 
     
     
         5 . The method of  claim 1 , wherein in at least one blueprint record, the target residue positions are in an order different from the order of target residues positions in the reference target structure. 
     
     
         6 . The method of  claim 1 , wherein the step of labeling the training set of blueprint comprises:
 performing computational protein modeling on that blueprint record to generate a polypeptide structure,   calculating a score for the polypeptide structure, and   associating the score with that blueprint record.   
     
     
         7 . The method of  claim 1 , wherein the computational protein modeling is based on a de novo design without template matching to the reference target structure. 
     
     
         8 . The method of  claim 1 , wherein each score in the labeling step comprises at least an energy term and a structure-constraint matching term that is determined using one or more structural constraints extracted from the representation of the reference target structure. 
     
     
         9 . The method of  claim 1 , further comprising:
 determining whether to retrain the machine learning model; and   if the machine learning model is to be retrained:   obtaining a second set of blueprint records;   generating a second training set of blueprint records from a predetermined portion of the reference target structure, wherein each blueprint record comprises one or more target residue positions and one or more scaffold residue positions, each target residue position corresponding to one target residue from the one or more target residues;   labeling each blueprint record of the second set training set of blueprint records with a second score; and   training a machine learning model based on the second training set.   
     
     
         10 . The method of  claim 9 , further comprising:
 concatenating, after the retraining the machine learning model, the first training set of blueprint records and the second training set of blueprint records to generate a retraining set of blueprint records and retraining scores, each blueprint record from the retraining blueprint records is associated with the retraining scores.   
     
     
         11 . The method of  claim 1 , wherein the at least one score is a preset value. 
     
     
         12 . The method of  claim 1 , wherein the at least one score is dynamically determined. 
     
     
         13 . The method of  claim 1 , wherein the machine learning model is a supervised machine learning model. 
     
     
         14 . The method of  claim 1 , comprising performing computational protein modeling of the output set of blueprint records to generate predicted structure of engineered polypeptides. 
     
     
         15 . The method of  claim 14 , comprising filtering the engineered polypeptides by static structure comparison to a representation of the reference target structure. 
     
     
         16 . The method of  claim 15 , comprising filtering the engineered polypeptides by dynamic structure comparison to a representation of the reference target structure using molecular dynamics (MD) simulations of the representation of the reference target structure and each of the structures of engineered polypeptides. 
     
     
         17 . A non-transitory processor-readable medium storing code representing instructions to be executed by a processor for designing engineered polypeptides using a machine learning model, the code comprising code to cause the processor to:
 receive a representation of a reference target structure for a reference target;   generate a training set of blueprint records from a predetermined portion of the reference target structure, wherein each blueprint record comprises one or more target residue positions and one or more scaffold residue positions, each target residue position corresponding to one target residue from the one or more target residues;   label each blueprint record of the training set of blueprint records with a score;   train a machine learning model based on the training set; and   apply the trained machine learning model to a set of desired scores to generate an output set of blueprint records with one or more desired scores, wherein the output set of blueprint records are used to design the engineered polypeptides.   
     
     
         18 . The non-transitory processor-readable medium of  claim 17  wherein the set of desired scores is dynamically determined. 
     
     
         19 . The non-transitory processor-readable medium of  claim 17 , wherein the machine learning model is a supervised machine learning model 
     
     
         20 . The non-transitory processor-readable medium of  claim 17 , comprising code to cause the processor to:
 perform computational protein modeling on the output set of blueprint records to generate predicted structures of engineered polypeptides.   
     
     
         21 . The non-transitory processor-readable medium of  claim 20 , comprising code to cause the processor to:
 filter the first predicted structures of the engineered polypeptides by static structure comparison to a representation of a reference target structure.   
     
     
         22 . The non-transitory processor-readable medium of  claim 17 , further comprising generating one or more polypeptides sequences from a blueprint record of the output set of blueprint records by computational protein modeling. 
     
     
         23 . The non-transitory processor-readable medium of  claim 17 , further comprising generating one or more engineered polypeptides having a polypeptide sequences generated from the output set of blueprint records by computational protein modeling. 
     
     
         24 . An apparatus for selecting an engineered polypeptide using a machine learning model, comprising:
 a first compute device having a processor and a memory storing instructions executable by the processor to:   receive a representation of a reference target structure for a reference target;   generate a training set of blueprint records from a predetermined portion of the reference target structure, wherein each blueprint record comprises one or more target residue positions and one or more scaffold residue positions, each target residue position corresponding to one target residue from the one or more target residues;   label each blueprint record of the training set of blueprint records with a score;   train a machine learning model based on the training set; and   apply the trained machine learning model to a set of desired scores to generate an output set of blueprint records with one or more desired scores, wherein the output set of blueprint records are used to design the engineered polypeptides.   
     
     
         25 . The apparatus of  claim 24 , wherein the set of desired scores is dynamically determined. 
     
     
         26 . The apparatus of  claim 24 , wherein the machine learning model is a supervised machine learning model. 
     
     
         27 . The apparatus of  claim 24 , comprising code to cause the processor to:
 perform computational protein modeling on the output set of blueprint records to generate predicted structures of engineered polypeptides.   
     
     
         28 . The apparatus of  claim 24 , comprising code to cause the processor to:
 filter the first predicted structures of the engineered polypeptides by static structure comparison to a representation of a reference target structure.   
     
     
         29 . The apparatus of  claim 24 , comprising code to cause the processor to:
 filter the engineered polypeptides by dynamic structure comparison to a representation of a reference target structure using molecular dynamics (MD) simulations of the representation of the reference target structure and each of the engineered polypeptides.   
     
     
         30 . The apparatus of  claim 24 , further comprising generating one or more polypeptides sequences from a blueprint record of the output set of blueprint records by computational protein modeling. 
     
     
         31 . The apparatus of  claim 24 , further comprising generating one or more engineered polypeptides having a polypeptide sequences generated from the output set of blueprint records by computational protein modeling. 
     
     
         32 . An engineered polypeptide generated by the method of  claim 1 .