Systems and methods for fully automated protein engineering
AssigneeWisconsin Alumni Research Foundation
InventorsPhilip Anthony Romero, Bennett James Bremer, Jacob Thomas Rapp
Life of this patent
Estimating the term…
Abstract
Systems and methods for protein engineering. The systems include a sequence testing subsystem and a machine learning subsystem. The sequence testing subsystem is configured to express proteins and test the expressed proteins for a given property. The machine learning subsystem is configured to model the activities of a set of possible proteins in light of the properties of the tested proteins and provide one or more untested proteins in the set to the sequence testing subsystem for subsequent testing. The system can be run in an iterative fashion and be fully automated. Methods of using the systems are provided.
Claims20
A protein engineering system comprising:
- a sequence testing subsystem comprising:
- nucleic acid fragments capable of being combined in different combinations to generate nucleic acids encoding a set of different possible proteins;
- a nucleic acid assembly subsystem that assembles the nucleic acids from the nucleic acid fragments;
- a protein expression subsystem that expresses proteins from the nucleic acids and thereby generates expressed proteins;
- and
- a property detector subsystem that detects property values of the expressed proteins for a given property and thereby generates detected property values corresponding to the expressed proteins;
- and
- a machine learning subsystem comprising a server that:
- receives the detected property values from the sequence testing subsystem;
- predicts, from the detected property values received from the sequence testing subsystem, predicted property values for untested proteins, wherein the untested proteins comprise possible proteins in the set that do not have corresponding detected property values;
- selects one or more of the untested proteins based on the predicted property values to thereby generate one or more selected proteins;
- and
- instructs the sequence testing subsystem to test the one or more selected proteins to obtain a detected property value therefor.
A method of protein engineering, comprising:
- in a sequence testing subsystem:
- assembling, in a nucleic acid assembly subsystem, nucleic acids from nucleic acid fragments to generate nucleic acids encoding a set of different possible proteins;
- expressing, in a protein expression subsystem, proteins from the nucleic acids to thereby generate expressed proteins;
- detecting, in a property detector subsystem, property values of the expressed proteins;
- and
- determining, in a machine learning subsystem, all possible nucleic acid sequences of the nucleic acids capable of being assembled from the nucleic acid fragments, thereby generating detected property values corresponding to the expressed proteins;
- and
- in a machine learning subsystem:
- receiving the detected property values from the sequence testing subsystem;
- predicting, from the detected property values received from the sequence testing subsystem, predicted property values for untested proteins, wherein the untested proteins comprise possible proteins in the set that do not have corresponding detected property values;
- selecting one or more of the untested proteins based on the predicted property values to thereby generate one or more selected proteins;
- and
- instructing the sequence testing subsystem to test the one or more selected proteins to obtain a detected property value therefor,
- wherein each of the assembling, the expressing, the detecting, the determining, the receiving, the predicting, the selecting, and the instructing are performed in an automated, computer-controlled system.
Dependencies are inferred from the claim text.