Predicting interactions between drugs and diseases

Patent metadata

Field	Value
Publication number	US-11276494-B2
Application number	US-201815976970-A
Country	US
Kind code	B2
Filing date	May 11, 2018
Priority date	May 11, 2018
Publication date	Mar 15, 2022
Grant date	Mar 15, 2022

Abstract

Embodiments of the present invention disclose a method, a computer program product, and a computer system for predicting drug and disease interactions. A computer identifies one or more drug similarity measures between one or more drugs and one or more disease similarity measures between one or more diseases. In addition, the computer identifies one or more interactions between the one or more drugs and the one or more diseases, then calculates one or more drug-disease feature vectors based on the one or more interactions, the one or more drug similarity measures, and the one or more disease similarity measures. Furthermore, the computer calculates a first probability indicating whether a first drug of the one or more drugs will interact with a first disease of the one or more diseases based on a model, wherein the model is trained based on the one or more drug-disease feature vectors.

First claim

The invention claimed is:

1. A method for predicting drug-disease interactions, the method comprising:
a computer generating a knowledge graph based on ingesting information detailing one or more drugs and one or more diseases;
the computer identifying one or more drug similarity measures between at least two of the one or more drugs based on the knowledge graph, wherein the one or more drug similarity measures include chemical-protein interactome profile and measures selected from the group comprising chemical structure, drug target, mechanism of action, anatomical therapeutic chemical, metabolizing enzyme, medical subject headings category, side effect, physiological effect, and pathway;
the computer identifying one or more disease similarity measures between at least two of the one or more diseases based on the knowledge graph, wherein the one or more disease similarity measures include disease pathology and measures selected from the group comprising semantic, disease ontology, disease gene, disease phenotype, disease symptom, disease comorbidity, disease morphology, application independent, domain specific, process based, molecular, and cellular;
the computer identifying one or more interactions between the one or more drugs and the one or more diseases;
the computer building one or more drug-disease pair feature vectors based on one or more similarity measures between the one or more interactions, the one or more drug similarity measures, and the one or more disease similarity measures, the one or more similarity measures selected from a group consisting of a max, a mean over positive pairs, a standard deviation over positive pairs, a max Z value, and a mean over all pairs;
the computer training one or more logistic regression models, wherein the one or more interactions are used as training data, and wherein the one or more drug-disease pair feature vectors are used as variables; and
the computer applying the one or more logistic regression models in order to calculate a first probability indicating whether a first drug of the one or more drugs will interact with a first disease of the one or more diseases.

2. The method of claim 1 , further comprising:
based on determining that the first probability exceeds a threshold, the computer identifying a second drug having a same intended result of the first drug; and
the computer determining a second probability indicating whether the second drug will interact with the first disease.

3. The method of claim 1 , further comprising:
utilizing the first probability in a causality assessment.

4. A computer program product for predicting drug-disease interactions, the computer program product comprising:
one or more computer-readable storage media and program instructions stored on the one or more computer-readable storage media, the program instructions comprising:
program instructions to generate a knowledge graph based on ingesting information detailing one or more drugs and one or more diseases;
program instructions to identify one or more drug similarity measures between at least two of the one or more drugs based on the knowledge graph, wherein the one or more drug similarity measures include chemical-protein interactome profile and measures selected from the group comprising chemical structure, drug target, mechanism of action, anatomical therapeutic chemical, metabolizing enzyme, medical subject headings category, side effect, physiological effect, and pathway;
program instructions to identify one or more disease similarity measures between at least two of the one or more diseases based on the knowledge graph, wherein the one or more disease similarity measures include disease pathology and measures selected from the group comprising semantic, disease ontology, disease gene, disease phenotype, disease symptom, disease comorbidity, disease morphology, application independent, domain specific, process based, molecular, and cellular;
program instructions to identify one or more interactions between the one or more drugs and the one or more diseases;
program instructions to build one or more drug-disease pair feature vectors based on one or more similarity measures between the one or more interactions, the one or more drug similarity measures, and the one or more disease similarity measures, the one or more similarity measures selected from a group consisting of a max, a mean over positive pairs, a standard deviation over positive pairs, a max Z value, and a mean over all pairs;
program instructions to train one or more logistic regression models, wherein the one or more interactions are used as training data, and wherein the one or more drug-disease pair feature vectors are used as variables; and
program instructions to apply the one or more logistic regression models in order to calculate a first probability indicating whether a first drug of the one or more drugs will interact with a first disease of the one or more diseases.

5. The computer program product of claim 4 , further comprising:
based on determining that the first probability exceeds a threshold, program instructions to identify a second drug having a same intended result of the first drug; and
program instructions to determine a second probability indicating whether the second drug will interact with the first disease.

6. The computer program product of claim 4 , further comprising:
program instructions to utilize the first probability in a causality assessment.

7. A computer system for predicting drug-disease interactions, the computer system comprising:
one or more computer processors, one or more computer-readable storage media, and program instructions stored on one or more of the computer-readable storage media for execution by at least one of the one or more processors, the program instructions comprising:
program instructions to generate a knowledge graph based on ingesting information detailing one or more drugs and one or more diseases;
program instructions to identify one or more drug similarity measures between at least two of the one or more drugs based on the knowledge graph, wherein the one or more drug similarity measures include chemical-protein interactome profile and measures selected from the group comprising chemical structure, drug target, mechanism of action, anatomical therapeutic chemical, metabolizing enzyme, medical subject headings category, side effect, physiological effect, and pathway;
program instructions to identify one or more disease similarity measures between at least two of the one or more diseases based on the knowledge graph, wherein the one or more disease similarity measures include disease pathology and measures selected from the group comprising semantic, disease ontology, disease gene, disease phenotype, disease symptom, disease comorbidity, disease morphology, application independent, domain specific, process based, molecular, and cellular;
program instructions to identify one or more interactions between the one or more drugs and the one or more diseases;
program instructions to build one or more drug-disease pair feature vectors based on one or more similarity measures between the one or more interactions, the one or more drug similarity measures, and the one or more disease similarity measures, the one or more similarity measures selected from a group consisting of a max, a mean over positive pairs, a standard deviation over positive pairs, a max Z value, and a mean over all pairs;
program instructions to train one or more logistic regression models, wherein the one or more interactions are used as training data, and wherein the one or more drug-disease pair feature vectors are used as variables; and
program instructions to apply the one or more logistic regression models in order to calcul

Assignees

• IBM

Inventors

• Zhang Ping
• Fokoue-Nkoutche Achille B
• Dey Sanjoy
• Shen Katherine

Classifications

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G16H50/70 Primary

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

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title

• G16H70/40

  relating to drugs, e.g. their side effects or intended usage · CPC title

• G16H50/20 Primary

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