Methods and apparatuses for prediction of mechanism of activity of compounds

Patent metadata

Field	Value
Publication number	US-2018372724-A1
Application number	US-201816019332-A
Country	US
Kind code	A1
Filing date	Jun 26, 2018
Priority date	Jun 26, 2017
Publication date	Dec 27, 2018
Grant date	—

Abstract

A platform configured to predict type or family of an unknown drug candidate compound, the platform including: a living cell or a tissue; a detector that measures an indicator of a cellular response by the living cell or tissue upon exposure to the unknown drug candidate compound; a memory configured to store data related to the indicator of the cellular response detected by the detector from a library of drug types and/or families; and one or more processing unit(s) configured to: process the data related to the indicator of the cellular response of the living cell or tissue upon exposure to the unknown drug candidate compound, and compare cellular response data from the library of drug types and/or families, so that a drug type and/or a drug family and/or a mechanism of action of the unknown drug candidate compound can be predicted on the basis of a similarity between the detected cellular response data of the unknown drug candidate compound and the cellular response data of the library of drug types and/or families. Also disclosed are methods of screening an unknown drug, including: comparing the data measured from a test cell to corresponding cellular response data in a library of known drug types, and determining a relationship between the unknown drug and a known drug type or a known drug family to predict the type or family of the unknown drug.

First claim

What is claimed is:

1 . A platform configured to predict type or family of an unknown drug candidate compound, the platform comprising:
(a) a living cell or a tissue that is capable of exerting a force in response to exposure to the unknown drug candidate compound; (b) a detector that measures an indicator of a cellular response by the living cell or tissue upon exposure to the unknown drug candidate compound; (c) a memory configured to store data related to the indicator of the cellular response detected by the detector from a library of drug types and/or families; and (d) one or more processing unit(s) configured to:
(i) process the data related to the indicator of the cellular response of the living cell or tissue upon exposure to the unknown drug candidate compound, and
(ii) compare cellular response data from the library of drug types and/or families, so that a drug type and/or a drug family and/or a mechanism of action of the unknown drug candidate compound can be predicted on the basis of a similarity between the detected cellular response data of the unknown drug candidate compound and the cellular response data of the library of drug types and/or families.

2 . The platform according to claim 1 , wherein the living cell or tissue is a model of cardiac muscle fiber.

3 . The platform according to claim 1 , wherein the living cell or tissue is configured as a human cardiac tissue strips (hCTS).

4 . The platform according to claim 1 , wherein the platform is configured to electrically pace the living cell or tissue and wherein the cellular response data is captured at a variety of pacing frequencies.

5 . The platform according to claim 1 , wherein the processing unit is configured to implement machine learning.

6 . The platform according to claim 5 , wherein the machine learning utilizes predetermined parameters of cellular response data to classify the cellular response data measured in response to the unknown drug and the cellular response data from the library of drug types and/or families.

7 . The platform according to claim 6 , wherein the predetermined parameters of the cellular response data comprise force data, the force data comprising one or more of the following parameters:
(a) pacing frequency; (b) a captured pacing frequency; (c) a maximum force generated (amplitude); (d) a duration of rise from a cutoff level to maximum force in a contraction phase; (e) a duration of decline from maximum force to a cutoff level in a relaxation phase; (f) an area under the curve of rise from a cutoff level to maximum force; (g) an area under the curve of decline from maximum force to a cutoff level; (h) a maximum change of force over time (ΔF/Δt) of contraction phase; and (i) a maximum change of force over time (ΔF/Δt) of relaxation phase.

8 . The platform according to claim 7 , wherein the predetermined parameters of the cellular response data comprise force data, the force data comprising one or more of the following parameters:
desired pacing frequency, captured pacing frequency, max force generated (amplitude), duration of rise from 95% cutoff to max force (contraction phase), duration of decline from max force to 95% cutoff (relaxation phase), area under the curve of rise from 95% cutoff to max force, area under the curve of decline from max force to 95% cutoff, max change of force over time (ΔF/Δt) of contraction phase, max change of force over time (ΔF/Δt) of relaxation phase, duration of rise from 50% cutoff to max force, duration of decline from max force to 50% cutoff, area under the curve of rise from 50% cutoff to max force, area under the curve of decline from max force to 50% cutoff, duration of rise from 25% cutoff to max force, duration of decline from max force to 25% cutoff to max force, area under the curve of rise from 50% cutoff to max force, and area under the curve of decline from max force to 50% cutoff.

9 . The platform according to claim 1 , wherein the cellular response data comprises a measure of cell or tissue motion and/or electrical conduction and/or calcium flux and the detector is capable of detecting motion and/or electrical conduction and/or calcium flux in the living cell or tissue following exposure to the drug.

10 . The platform according to claim 9 , wherein the electrical conduction detected corresponds to one or more of a micro-impedance signal and an electrophysiological signal.

11 . The platform according to claim 1 , wherein the processing unit is configured to output dosing information of the unknown drug candidate compound based upon a comparison to the cellular response data of one or more members of the library.

12 . The platform according to claim 1 , further comprising a library of drug types and/or families stored in the memory.

13 . The platform according to claim 12 , wherein each drug type or drug family is characterized by a plurality of distinct compounds within the drug type or drug family.

14 . A method of screening an unknown drug, comprising:
(a) exposing a test cell or a tissue to the unknown drug, (b) quantifying a cellular response by obtaining cellular response data measured from the test cell in response to the unknown drug, (c) comparing the data measured from the test cell to corresponding cellular response data in a library of known drug types, (d) determining a relationship between the unknown drug and a known drug type or a known drug family to predict the type or family of the unknown drug.

15 . The method of claim 14 , wherein the cellular response data is indicative of cardioactivity.

16 . The method of claim 14 wherein the test cell or tissue is a human cardiac tissue construct.

17 . The method according to claim 15 , wherein a degree to which a compound is cardiotoxic/cardioactive is predicted.

18 . The method according to claim 14 , wherein machine learning is used to form the library of cellular response data of known drug types.

19 . The method according to claim 14 , wherein said comparing the cellular response data of the test cell to a library of corresponding cellular response data of known drug types is done by a series of binary classifications.

20 . The method according to claim 14 comprising calculating a singular quantitative index generated by a supervised learning algorithm to consolidate a plurality of parameters of a cellular response into a singular quantitative index.

Assignees

• Univ California
• Novoheart Ltd

Inventors

• Khine Michelle
• Lee Eugene
• Li Tang Wai Ronald Adolphus
• Tran David Dan

Classifications

• G06F19/24

  Physics · mapped topic

• G01N33/5061 Primary

  Muscle cells · CPC title

• G16C20/70

  Machine learning, data mining or chemometrics · CPC title

• G16C20/64

  Screening of libraries · CPC title

• G16C20/40

  Searching chemical structures or physicochemical data · CPC title