Individual and cohort pharmacological phenotype prediction platform

• US 10,249,389 B2
• Filed: 05/11/2018
• Issued: 04/02/2019
• Est. Priority Date: 05/12/2017
• Status: Active Grant

• Current Assignee: Board of Regents of the University of Michigan (University of Michigan)
• Original Assignee: Board of Regents of the University of Michigan (University of Michigan)
• Inventors: Athey, Brian D., Allyn-Feuer, Ari, Higgins, Gerald A., Burns, James S., Kalinin, Alexandr, Pauls, Brian, Ade, Alex, Reamaroon, Narathip
• Primary Examiner(s): Pope, Daryl C
• Application Number:
  US15/977,347
• Publication Number:
  US 20180330824A1
• Time in Patent Office: 326 Days
• Field of Search: 340506, 340 31, 3405391, 34053911, 34053913, 34053912, 3405731
• CPC Class Codes: A61K 31/37 Coumarins, e.g. psoralen
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G16B 20/00 ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
• CPC Class Codes: G16B 20/20 Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
• CPC Class Codes: G16B 30/00 ICT specially adapted for sequence analysis involving nucleotides or amino acids
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
• CPC Class Codes: G16B 40/20 Supervised data analysis
• CPC Class Codes: G16B 40/30 Unsupervised data analysis
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

For patients who exhibit or may exhibit primary or comorbid disease, pharmacological phenotypes may be predicted through the collection of panomic, physiomic, environmental, sociomic, demographic, and outcome phenotype data over a period of time. A machine learning engine may generate a statistical model based on training data from training patients to predict pharmacological phenotypes, including drug response and dosing, drug adverse events, disease and comorbid disease risk, drug-gene, drug-drug, and polypharmacy interactions. Then the model may be applied to data for new patients to predict their pharmacological phenotypes, and enable decision making in clinical and research contexts, including drug selection and dosage, changes in drug regimens, polypharmacy optimization, monitoring, etc., to benefit from additional predictive power, resulting in adverse event and substance abuse avoidance, improved drug response, better patient outcomes, lower treatment costs, public health benefits, and increases in the effectiveness of research in pharmacology and other biomedical fields.

Claims

1. A computer-implemented method for identifying pharmacological phenotypes using statistical modeling and machine learning techniques, the method executed by one or more processors programmed to perform the method, the method comprising:

obtaining, at one or more processors, a set of training data including for each of a plurality of first patients;

panomic data indicative of biological characteristics of the first patient,

sociomic data indicative of risk factors associated with adverse cultural, childhood, acute or chronic traumatic events, or chronic stress resulting from adverse conditions,

environmental data indicative of experiences of the first patient collected over time, and

phenomic data indicative of at least one of;

a response to one or more drugs, whether the first patient experiences substance abuse, or one or more chronic diseases of the first patient;

generating, by the one or more processors, a statistical model for determining pharmacological phenotypes based on the set of training data;

receiving, at the one or more processors, a set of panomic data, and sociomic and environmental data for a second patient collected over a period of time;

applying, by the one or more processors, the panomic data, and the sociomic and environmental data for the second patient to the statistical model to determine one or more pharmacological phenotypes for the second patient; and

providing, by the one or more processors, the one or more pharmacological phenotypes for the second patient for display to a health care provider, wherein the health care provider recommends a course of treatment to the second patient according to the pharmacological phenotypes.

2. The method of claim 1, wherein the sociomic and environmental data is indicative of experiences of the first patient collected over time includes at least one of:

clinical data indicative of a medical history of the first patient, demographic data of the first patient, socioeconomic data of the first patient, or polypharmacy data indicative of each drug prescribed to the first patient collected from at least two instances of time.

3. The method of claim 1, wherein the one or more pharmacological phenotypes for the second patient include at least one of:

a predicted response by the second patient to one or more drugs, a risk of one or more illnesses, an adverse drug event or adverse drug reaction to the one or more drugs, or a likelihood of substance abuse.

4. The method of claim 1, wherein the one or more pharmacological phenotypes includes a predicted response by the second patient to one or more drugs and further comprising:

receiving, at the one or more processors, an indication of a type of disease diagnosed for the second patient;

applying, by the one or more processors, the panomic data and the sociomic and environmental data for the second patient to the statistical model to determine the predicted response by the second patient to a plurality of drugs for treating the type of disease diagnosed for the second patient;

identifying, the one or more processors, one of the plurality of drugs to treat the second patient with based on the predicted response for each of the plurality of prescriptions drugs; and

providing, by the one or more processors, an indication of the identified drug to the health care provider for prescribing an optimal course of treatment for the type of disease diagnosed for the second patient.

5. The method of claim 4, wherein identifying one of the plurality of drugs to treat the second patient with based on the predicted response includes:

for each of the plurality of drugs;

determining, by the one or more processors, an efficacy of the drug with respect to the second patient based on the panomic data and the sociomic and environmental data for the second patient;

determining, by the one or more processors, one or more expected adverse drug reactions with respect to the second patient based on the panomic data and the sociomic and environmental data for the second patient;

combining, by the one or more processors, the efficacy and the one or more expected adverse drug reactions to determine an overall value of the drug to the second patient; and

identifying, by the one or more processors, the drug of the plurality of drugs having the highest overall value for the second patient.

6. The method of claim 1, wherein the one or more pharmacological phenotypes includes a likelihood of substance abuse and further comprising:

applying, by the one or more processors, the panomic data and the sociomic and environmental data for the second patient to the statistical model to determine that the second patient has a likelihood of substance abuse that exceeds a threshold likelihood; and

providing, by the one or more processors, an indication that the second patient is expected to suffer from substance abuse problems for display to a health care provider for early intervention.

7. The method of claim 1, wherein providing the one or more pharmacological phenotypes for the second patient for display to a health care provider includes:

generating, by the one or more processors, a risk analysis display for the second patient including at least one of;

a predicted response by the second patient to each of one or more drugs, a risk of one or more illnesses, or a likelihood of substance abuse.

8. The method of claim 1, further comprising:

after a threshold time period, receiving, by the one or more processors, phenomic data for the second patient indicative of at least one of;

a response to one or more drugs, whether the second patient experiences substance abuse, or one or more chronic diseases of the second patient;

storing, by the one or more processors in a knowledge base, the panomic data, the sociomic and environmental data, and the phenomic data for the second patient; and

updating, by the one or more processors, the set of training data to include data from the knowledge base.

9. The method of claim 1, wherein panomic data includes at least one of genomic data, epigenomic data, transcriptomic data, proteomic data, choromosomic data, or metabolomic data.

10. The method of claim 1, wherein the statistical model is generated using one or more machine learning techniques.

11. The method of claim 1, wherein generating the statistical model for determining pharmacological phenotypes includes identifying panomic data that corresponds to a set of pharmacological phenotypes, including:

identifying a plurality of single nucleotide polymorphisms (SNPs) correlated with the set of pharmacological phenotypes;

comparing the plurality of SNPs to a database of SNPs to identify additional SNPs that are linked to the plurality of SNPs, wherein the plurality of SNPs and additional SNPs are included in a set of permissive candidate variants;

performing a bioinformatics analysis to filter the set of permissive candidate variants into a subset of intermediate candidate variants based on at least one of;

regulatory function, variant dependence, a presence of target gene relationships for the permissive candidate variants, or whether the permissive candidate variants are non-synonymous coding variants with a minor allele frequency;

ranking the subset of intermediate candidate variants;

determining intermediate candidate variants of the subset that are ranked above a threshold ranking to identify SNPs, genes associated with the SNPs, or pathways associated with the SNPs which are causally related to the set of pharmacological phenotypes.

12. The method of claim 11, wherein panomic data is identified that corresponds to a set of warfarin phenotypes, and wherein the identified genes or pathways which are causally related to the set of warfarin phenotypes include a warfarin response pathway.

13. The method of claim 12, wherein the warfarin response pathway includes one or more of:

complement factor H related 5 (CFHR5) gene, fibrinogen alpha chain gene (FGA) gene, flavin containing monoxygenase 5 (FMO5) gene, histidine rich glycoprotein (HRG) gene, kininogen 1 (KNG1) gene, surfeit 4 (SURF4) gene, alpha 1-3-N-acetylgalactosaminyltransferase and alpha 1-3-galactosyltransferase (ABO) gene, lysozyme (LYZ) gene, polycomb group ring finger 3 (PCGF3) gene, serine protease 8 (PRSS8) gene, transient receptor potential cation channel subfamily C member 4 associated pattern (TRPC4AP) gene, solute carrier family 44 member 2 (SLC44A2) gene, sphingosine kinase 1 (SPHK1) gene, or ubiquitin specific peptidase 7 (USP7) gene.

14. The method of claim 12, wherein the set of warfarin phenotypes include one or more of:

coagulation, bleeding, clotting, thrombosis, or hemorrhaging.

15. The method of claim 12, wherein applying the panomic data, and the sociomic and environmental data for the second patient to the statistical model to determine one or more pharmacological phenotypes for the second patient includes comparing the panomic data for the second patient to genes or SNPs within the warfarin response pathway to determine one or more warfarin phenotypes for the second patient.

16. A computing device for identifying pharmacological phenotypes using statistical modeling and machine learning techniques, the computing device comprising:

a communication network,

one or more processors; and

a non-transitory computer-readable memory coupled to the one or more processors and storing thereon instructions that, when executed by the one or more processors, cause the computing device to;

obtain a set of training data including for each of a plurality of first patients;

panomic data indicative of biological characteristics of the first patient, sociomic and environmental data indicative of experiences of the first patient collected over time, and

phenomic data indicative of at least one of;

a response to one or more drugs, whether the first patient experiences substance abuse, or one or more chronic diseases of the first patient;

generate a statistical model for determining pharmacological phenotypes s based on the set of training data;

receive a set of panomic data, and sociomic and environmental data for a second patient collected over a period of time;

apply the panomic data and the sociomic and environmental data for the second patient to the statistical model to determine one or more pharmacological phenotypes for the second patient; and

provide, via the communication network, the one or more pharmacological phenotypes for the second patient for display to a health care provider, wherein the health care provider recommends a course of treatment to the second patient according to the pharmacological phenotypes.

17. The computing device of claim 16, wherein the sociomic and environmental data is indicative of experiences of the first patient collected over time includes at least one of:

clinical data indicative of a medical history of the first patient, demographic data of the first patient, socioeconomic data of the first patient, or polypharmacy data indicative of each drug prescribed to the first patient collected from at least two instances of time.

18. The computing device of claim 16, wherein the one or more pharmacological phenotypes for the second patient include at least one of:

a predicted response by the second patient to one or more drugs, a risk of one or more illnesses, an adverse drug event or adverse drug reaction to the one or more drugs, or a likelihood of substance abuse.

19. The computing device of claim 16, wherein the one or more pharmacological phenotypes includes a predicted response by the second patient to one or more drugs and the instructions further cause the computing device to:

receive an indication of a type of disease diagnosed for the second patient;

apply the panomic data and the sociomic and environmental data for the second patient to the statistical model to determine the predicted response by the second patient to a plurality of drugs for treating the type of disease diagnosed for the second patient;

identify one of the plurality of drugs to treat the second patient with based on the predicted response for each of the plurality of prescriptions drugs; and

provide an indication of the identified drug to the health care provider for prescribing an optimal course of treatment for the type of disease diagnosed for the second patient.

20. The computing device of claim 19, wherein to identify one of the plurality of drugs to treat the second patient with based on the predicted response, the instructions cause the computing device to:

for each of the plurality of drugs;

determine an efficacy of the drug with respect to the second patient based on the panomic data and the sociomic and environmental data for the second patient;

determine one or more expected adverse drug reactions with respect to the second patient based on the panomic data and the sociomic and environmental data for the second patient;

combine the efficacy and the one or more expected adverse drug reactions to determine an overall value of the drug to the second patient; and

identify the drug of the plurality of drugs having the highest overall value for the second patient.

21. The computing device of claim 16, wherein the one or more pharmacological phenotypes includes a likelihood of stress-related disorders, neuropsychiatric disorders, anxiety disorders, pain, substance abuse, cardiac disorders, or gastrointestinal disorders, and the instructions cause the computing device to:

apply the panomic data and the sociomic and environmental data for the second patient to the statistical model to determine that the second patient has a likelihood of stress-related disorders, neuropsychiatric disorders, anxiety disorders, pain, substance abuse, cardiac disorders, or gastrointestinal disorders that exceeds a threshold likelihood; and

provide an indication that the second patient is expected to suffer from stress-related disorders, neuropsychiatric disorders, anxiety disorders, pain, substance abuse, cardiac disorders, or gastrointestinal disorders for display to a health care provider for early intervention.

22. The computing device of claim 16, wherein to provide the one or more pharmacological phenotypes for the second patient for display to a health care provider, the instructions cause the computing device to:

generate a risk analysis display for the second patient including at least one of;

a predicted response by the second patient to each of one or more drugs, a risk of one or more illnesses, or a likelihood of substance abuse.

23. The computing device of claim 16, wherein the instructions further cause the computing device to:

after a threshold time period, receive phenomic data for the second patient indicative of at least one of;

a response to one or more drugs, whether the second patient experiences substance abuse, or one or more chronic diseases of the second patient;

store, in a knowledge base, the panomic data, the sociomic and environmental data, and the phenomic data for the second patient; and

update the set of training data to include data from the knowledge base.

24. The computing device of claim 16, wherein panomic data includes at least one of genomic data, epigenomic data, transcriptomic data, proteomic data, choromosomic data, or metabolomics data.

25. The computing device of claim 16, wherein the statistical model is generated using one or more machine learning techniques.