Back to results
US 11,599,800 B2Granted
Systems and methods for enhanced user specific predictions using machine learning techniques
AssigneeColor Genomics, Inc.
InventorsCarmen Lai, Jill Hagenkord, Katsuya Noguchi
Life of this patent
Estimating the term…
Abstract
Data sets can be processed using machine learning or artificial intelligence models to generate outputs predictive of a degree to which performing a protocol can positively modify an expected result associated with a condition. Generating the output may include accessing a user data set, inputting the user data set into a trained machine learning model to generate an output, and selecting an incomplete subset of a set of genes based on the output.
Claims23
A computer-implemented method, comprising:
- receiving, at a server, a first communication corresponding to a request to identify a gene panel that is specific to a user;
- in response to receiving the first communication, generating a second communication responsive to the request to identify the gene panel that is specific to the user, wherein identifying the gene panel includes:
- identifying a set of genes, each gene of the set of genes being associated with one or more diseases, and each disease of the one or more diseases being associated with one or more treatments;
- accessing a user data set associated with the user, the user data set including two or more user feature values, each user feature value of the two or more user feature values characterizing an aspect of the user;
- generating an input data set that includes the user data set;
- inputting the input data set into a trained machine learning model, the trained machine learning model having been trained using a training data set associated with a plurality of other users;
- processing the input data set using the trained machine learning model, wherein the processing includes executing the machine learning model multiple times, and wherein each of the executions uses user-specific data and wherein each of the executions generates:
- a set of intermediate values, each intermediate value of the set of intermediate values representing a computed value determined using:
- a set of real user attributes from the input data set and at least one estimated user attribute;
- and
- at least one parameter of the trained machine learning model, wherein at least some of the set of intermediate values corresponds to the set of genes, wherein the computed value represented by one or more of the set of intermediate values is determined based on the at least one estimated user attribute that corresponds to a user attribute that is missing from the user data set;
- and
- an output corresponding to a prediction of whether or a degree to which an expected event of a particular disease of the one or more diseases is modifiable performing, on the user, a particular treatment of the one or more treatments;
- determining, based at least in part on the executions of the machine learning model, that one or more particular user attributes are predictive of an output corresponding to a personalized prediction that treating the user with the particular treatment will positively modify the particular disease;
- and
- selecting a subset of genes from amongst the set of genes to include in the gene panel specific to the user based on the one or more particular user attributes;
- and
- transmitting the second communication in response to receiving the first communication.
15 more elements
A system, comprising:
- one or more processors;
- and
- a non-transitory computer-readable storage medium containing instructions which, when executed on the one or more processors, cause the one or more processors to perform operations including:
- receiving, at a server, a first communication corresponding to a request to identify a gene panel that is specific to a user;
- in response to receiving the first communication, generating a second communication responsive to the request to identify the gene panel that is specific to the user, wherein identifying the gene panel includes:
- identifying a set of genes, each gene of the set of genes being associated with one or more diseases, and each disease of the one or more diseases being associated with one or more treatments;
- accessing a user data set associated with the user, the user data set including two or more user feature values, each user feature value of the two or more user feature values characterizing an aspect of the user;
- generating an input data set that includes the user data set;
- inputting the input data set into a trained machine learning model, the trained machine learning model having been trained using a training data set associated with a plurality of other users;
- processing the input data set using the trained machine learning model, wherein the processing includes executing the machine learning model multiple times, and wherein each of the executions uses user-specific data and wherein each of the executions generates:
- a set of intermediate values, each intermediate value of the set of intermediate values representing a computed value determined using:
- a set of real user attributes from the input data set and at least one estimated user attribute;
- and
- at least one parameter of the trained machine learning model, wherein at least some of the set of intermediate values corresponds to the set of genes, wherein the computed value represented by one or more of the set of intermediate values is determined based on the at least one estimated user attribute that corresponds to a user attribute that is missing from the user data set;
- and
- an output corresponding to a prediction of whether or a degree to which an expected event of a particular disease of the one or more diseases is modifiable performing, on the user, a particular treatment of the one or more treatments;
- determining, based at least in part on the executions of the machine learning model, that one or more particular user attributes are predictive of an output corresponding to a personalized prediction that treating the user with the particular treatment will positively modify the particular disease;
- and
- selecting a subset of genes from amongst the set of genes to include in the gene panel specific to the user based on the one or more particular user attributes;
- and
- transmitting the second communication in response to receiving the first communication.
18 more elements
A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause a processing apparatus to perform operations including:
- receiving, at a server, a first communication corresponding to a request to identify a gene panel that is specific to a user;
- in response to receiving the first communication, generating a second communication responsive to the request to identify the gene panel that is specific to the user, wherein identifying the gene panel includes:
- identifying a set of genes, each gene of the set of genes being associated with one or more diseases, and each disease of the one or more diseases being associated with one or more treatments;
- accessing a user data set associated with the user, the user data set including two or more user feature values, each user feature value of the two or more user feature values characterizing an aspect of the user;
- generating an input data set that includes the user data set;
- inputting the input data set into a trained machine learning model, the trained machine learning model having been trained using a training data set associated with a plurality of other users;
- processing the input data set using the trained machine learning model, wherein the processing includes executing the machine learning model multiple times, and wherein each of the executions uses user-specific data and wherein each of the executions generates:
- a set of intermediate values, each intermediate value of the set of intermediate values representing a computed value determined using:
- a set of real user attributes from the input data set and at least one estimated user attribute;
- and
- at least one parameter of the trained machine learning model, wherein at least some of the set of intermediate values corresponds to the set of genes, wherein the computed value represented by one or more of the set of intermediate values is determined based on the at least one estimated user attribute that corresponds to a user attribute that is missing from the user data set;
- and
- an output corresponding to a prediction of whether or a degree to which an expected event of a particular disease of the one or more diseases is modifiable performing, on the user, a particular treatment of the one or more treatments;
- determining, based at least in part on the executions of the machine learning model, that one or more particular user attributes are predictive of an output corresponding to a personalized prediction that treating the user with the particular treatment will positively modify the particular disease;
- and
- selecting a subset of genes from amongst the set of genes to include in the gene panel specific to the user based on the one or more particular user attributes;
- and
- transmitting the second communication in response to receiving the first communication.
15 more elements
Dependencies are inferred from the claim text.