Predictive assignments that relate to genetic information and leverage machine learning models
Inventors
Glode, Christopher M. • Trunck, Ryan P. • Powers, Rani K. • Lescallett, Jennifer L.
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-10733509-B2
Publication Date
2020-08-04
Expiration Date
Abstract
Systems and methods are provided for performing predictive assignments pertaining to genetic information. One embodiment is a system that includes a genetic prediction server. The genetic prediction server includes an interface that acquires records that each indicate one or more genetic variants determined to exist within an individual, and a controller. The controller selects one or more machine learning models that utilize the genetic variants as input, and loads the machine learning models. For each individual in the records: the controller predictively assigns at least one characteristic to that individual by operating the machine learning models based on at least one genetic variant indicated in the records for that individual. The controller also generates a report indicating at least one predictively assigned characteristic for at least one individual, and transmits a command via the interface for presenting the report at a display.
Core Innovation
The invention describes a genetic prediction server and related method that predictively assigns genetic variants to individuals by using machine learning models. An interface acquires records that each indicate one or more characteristics determined to exist for an individual, and a controller selects one or more machine learning models that utilize characteristics as input to predictively assign at least one genetic variant using at least one characteristic indicated in the records as input.
The controller analyzes input indicating accuracy of a predictively assigned genetic variant and determines a score for a machine learning model based on the analyzed input and a cost function. Each machine learning model comprises a multi-layer neural network with multiple nodes per layer, where nodes in different layers are coupled via weighted connections, and nodes in an input layer correspond with characteristics and nodes in an output layer correspond with genetic variants.
The controller revises the weighted connections based on the score. The approach includes forward prediction by using characteristics to predict genetic variants, and it can also be applied in a reverse process where genetic variants are predicted from characteristics. The system generates and transmits reports for display and may use confidence values and confidence thresholds to decide whether to assign predicted genetic variants.
Claims Coverage
Independent claims are present in clm-00001, clm-00008, and clm-00015. The claims cover three inventive features centered on multi-layer neural network models that take characteristics as input to predictively assign genetic variants, score model accuracy with a cost function, and revise weighted connections based on the score.
Genetic prediction server with neural-network variant assignment and cost-function weight revision
A genetic prediction server with an interface that acquires records indicating characteristics for individuals, and a controller that selects one or more machine learning models using characteristics as input to predictively assign at least one genetic variant to each individual; the controller analyzes input indicating accuracy, determines a score based on the analyzed input and a cost function, where each model is a multi-layer neural network with weighted connections and an input layer corresponding to characteristics and an output layer corresponding to genetic variants, and revises the weighted connections based on the score.
Method of neural-network genetic variant assignment with cost-function scoring and weight revision
A method that acquires records indicating characteristics for an individual, selects one or more machine learning models that utilize characteristics as input, and for each individual predictively assigns at least one genetic variant by operating the one or more machine learning models using at least one characteristic indicated in the records as input; the method analyzes input indicating accuracy, determines a score based on the analyzed input and a cost function, where each model comprises a multi-layer neural network with weighted connections and an input layer corresponding to characteristics and an output layer corresponding to genetic variants, and revises the weighted connections based on the score.
Non-transitory computer readable medium for neural-network genetic variant assignment with cost-function weight revision
A non-transitory computer readable medium embodying programmed instructions operable to perform a method that acquires records indicating characteristics for an individual, selects one or more machine learning models that utilize characteristics as input, and for each individual predictively assigns at least one genetic variant by operating the one or more machine learning models using at least one characteristic indicated in the records as input; the method analyzes input indicating accuracy, determines a score for a machine learning model based on the analyzed input and a cost function, where each model comprises a multi-layer neural network with weighted connections and an input layer corresponding to characteristics and an output layer corresponding to genetic variants, and revises the weighted connections based on the score.
Across clm-00001, clm-00008, and clm-00015, the claims focus on multi-layer neural network models that take characteristics as input and output genetic variants, with model accuracy analyzed to determine a score via a cost function and with revision of the neural network’s weighted connections based on the score. Dependent claims refine the variant/phenotype relationship, add location-based characteristic representations, and include confidence-threshold logic for variant assignment.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?