Quantum computing techniques for generating epistatic polygenic risk scores

Patent metadata

Field	Value
Publication number	US-2022367058-A1
Application number	US-202117318479-A
Country	US
Kind code	A1
Filing date	May 12, 2021
Priority date	May 12, 2021
Publication date	Nov 17, 2022
Grant date	—

Abstract

Various embodiments of the present invention utilize systems, methods, and computer program products that perform health-related predictive data analysis by utilizing an epistatic polygenic risk score generation machine learning model comprises at least one of the following: (i) an epistatic interaction score generation sub-model that is configured to process one or more significant epistatic interaction features for the patient data object that correspond to one or more significant epistatic interactions defined by the epistatic interaction score generation sub-model in order to generate an epistatic interaction score, and (ii) a base polygenic risk score generation sub-model that is configured to process one or more significant genetic variant features for the patient data object that correspond to one or more significant genetic variants defined by the base polygenic risk score generation machine learning model in order to generate a base polygenic risk score.

First claim

1 . A computer-implemented method for generating an epistatic polygenic risk score, the computer-implemented method comprising:
identifying, using one or more processors, an epistatic polygenic risk score generation machine learning model, wherein:
the epistatic polygenic risk score generation machine learning model comprises an epistatic polygenic risk score generation sub-model,
the epistatic polygenic risk score generation sub-model is characterized by one or more significant epistatic interactions, and
the one or more significant epistatic interactions are determined by: (i) performing a set of quantum superposition operations on a genomic scan data object to generate a group of genome scan qubits, and (ii) performing a set of constructive interference determination operations across the group of genome scan qubits to determine the one or more significant epistatic interactions;
generating, using the one or more processors and the epistatic polygenic risk score generation machine learning model, the epistatic polygenic risk score, wherein generating the epistatic polygenic risk score using the epistatic polygenic risk score generation machine learning model comprises: (i) processing one or more significant epistatic interaction features corresponding to the one or more significant epistatic interactions using the epistatic polygenic risk score generation sub-model to generate an epistatic interaction score, and (ii) generating the epistatic polygenic risk score based at least in part on the epistatic interaction score, and performing, using the one or more processors, one or more prediction-based actions based at least in part on the epistatic polygenic risk score.

2 . The computer-implemented method of claim 1 , wherein the epistatic polygenic risk score generation sub-model defines an interaction weight for each significant epistatic interaction.

3 . The computer-implemented method of claim 2 , wherein each interaction weight for a significant epistatic interaction is determined by generating a maximum likelihood estimation for the significant epistatic interaction using a maximum likelihood estimator for a normally distributed nonlinear implicit relationship defined by the epistatic polygenic risk score generation machine learning model.

4 . The computer-implemented method of claim 1 , wherein the epistatic polygenic risk score generation machine learning model further comprises a base polygenic risk score generation sub-model that is characterized by one or more significant genetic variants.

5 . The computer-implemented method of claim 4 , wherein generating the epistatic polygenic risk score using the epistatic polygenic risk score generation machine learning model further comprises:
processing one or more significant genetic variant features corresponding to the one or more significant genetic variants using the base polygenic risk score generation sub-model to generate a base polygenic risk score, and generating the epistatic polygenic risk score based at least in part on the base polygenic risk score and the epistatic interaction score.

6 . The computer-implemented method of claim 1 , wherein performing the set of quantum superposition operations comprises:
identifying a group of loci associated with the genomic scan data object; for each locus, determining, based at least in part on applying a Hadamard quantum logic gate transformation of a nucleobase feature described by the locus, a superposed representation; and determining the group of genome scan qubits based at least in part on each superposed representation.

7 . The computer-implemented method of claim 1 , wherein the set of constructive interference determination operations are performed based at least in part on a maximal epistatic interaction order hyper-parameter of the epistatic polygenic risk score generation machine learning model.

8 . An apparatus for generating an epistatic polygenic risk score, the apparatus comprising at least one processor and at least one memory including program code, the at least one memory and the program code configured to, with the processor, cause the apparatus to at least:
identify an epistatic polygenic risk score generation machine learning model, wherein:
the epistatic polygenic risk score generation machine learning model comprises an epistatic polygenic risk score generation sub-model,
the epistatic polygenic risk score generation sub-model is characterized by one or more significant epistatic interactions, and
the one or more significant epistatic interactions are determined by: (i) performing a set of quantum superposition operations on a genomic scan data object to generate a group of genome scan qubits, and (ii) performing a set of constructive interference determination operations across the group of genome scan qubits to determine the one or more significant epistatic interactions;
generate, using the epistatic polygenic risk score generation machine learning model, the epistatic polygenic risk score, wherein generating the epistatic polygenic risk score using the epistatic polygenic risk score generation machine learning model comprises: (i) processing one or more significant epistatic interaction features corresponding to the one or more significant epistatic interactions using the epistatic polygenic risk score generation sub-model to generate an epistatic interaction score, and (ii) generating the epistatic polygenic risk score based at least in part on the epistatic interaction score, and perform one or more prediction-based actions based at least in part on the epistatic polygenic risk score.

9 . The apparatus of claim 8 , wherein the epistatic polygenic risk score generation sub-model defines an interaction weight for each significant epistatic interaction.

10 . The apparatus of claim 9 , wherein each interaction weight for a significant epistatic interaction is determined by generating a maximum likelihood estimation for the significant epistatic interaction using a maximum likelihood estimator for a normally distributed nonlinear implicit relationship defined by the epistatic polygenic risk score generation machine learning model.

11 . The apparatus of claim 8 , wherein the epistatic polygenic risk score generation machine learning model further comprises a base polygenic risk score generation sub-model that is characterized by one or more significant genetic variants.

12 . The apparatus of claim 11 , wherein generating the epistatic polygenic risk score using the epistatic polygenic risk score generation machine learning model further comprises:
processing one or more significant genetic variant features corresponding to the one or more significant genetic variants using the base polygenic risk score generation sub-model to generate a base polygenic risk score, and generating the epistatic polygenic risk score based at least in part on the base polygenic risk score and the epistatic interaction score.

13 . The apparatus of claim 8 , wherein performing the set of quantum superposition operations comprises:
identifying a group of loci associated with the genomic scan data object; for each locus, determining, based at least in part on applying a Hadamard quantum logic gate transformation of a nucleobase feature described by the locus, a superposed representation; and determining the group of genome scan qubits based at least in part on each superposed representation.

14 . The apparatus of claim 8 , wherein the set of constructive interference determination operations are performed based at least in part on a maximal epistatic interaction order hyper-parameter of the epistatic polyg

Assignees

• Optum Inc

Inventors

• Versaggi Matthew R
• Godden Paul J

Classifications

• G16H50/30 Primary

  for calculating health indices; for individual health risk assessment · CPC title

• G16H50/70 Primary

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G06N10/00

  Quantum computing, i.e. information processing based on quantum-mechanical phenomena · CPC title

• G06N20/00

  Machine learning · CPC title