Methods, computer-accessible medium and systems to model disease progression using biomedical data from multiple patients

Patent metadata

Field	Value
Publication number	US-11205519-B2
Application number	US-201816192418-A
Country	US
Kind code	B2
Filing date	Nov 15, 2018
Priority date	Oct 28, 2013
Publication date	Dec 21, 2021
Grant date	Dec 21, 2021

Abstract

An exemplary embodiment of system, method and computer-accessible medium can be provided to reconstruct models based on the probabilistic notion of causation, which can differ fundamentally from that can be based on correlation. A general reconstruction setting can be complicated by the presence of noise in the data, owing to the intrinsic variability of biological processes as well as experimental or measurement errors. To gain immunity to noise in the reconstruction performance, it is possible to use a shrinkage estimator. On synthetic data, the exemplary procedure can outperform currently known procedures and, for some real cancer datasets, there are biologically significant differences revealed by the exemplary reconstructed progressions. The exemplary system, method and computer accessible medium can be efficient even with a relatively low number of samples and its performance quickly converges to its asymptote as the number of samples increases.

First claim

What is claimed is:

1. A non-transitory computer-accessible medium having stored thereon computer-executable instructions for generating a model of progression of cancer using biomedical data of at least one patient, wherein, when a computer arrangement executes the instructions, the computer arrangement is configured by the instruction to perform procedures comprising:
obtaining the biomedical data which includes at least one of genomics, transcriptomics, or epigenomics;
generating the model of progression which includes (i) states of the cancer and (ii) transitions among the states using a learning network, based on the obtained biomedical data, wherein:
transitions among the states are determined by a causality relationship information whose strength is estimated by probability-raising by at least one estimator which is part of the learning network;
the at least one estimator includes at least one shrinkage estimator which is a measure of causation among any pair of events atomic events; and
the at least one shrinkage estimator is defined as θ=(1−λ)α(x)+λβ(x), where 0≤λ≤1 can be a shrinkage coefficient, x is an input data, and θ is one or more estimates for an evaluation, wherein the learning network includes the transitions and the at least one shrinkage estimator;

receiving further biomedical data associated with a further patient;
classifying a particular state of cancer for the further patient using the model of progression and the further biomedical data, wherein a particular state of the states of cancer is classified by one or more mutational profiles of the at least one of the genomics, the transcriptomics or the epigenomics; and
determining a genome-specific therapy design based on the classification of the particular state of cancer.

2. The computer-accessible medium of claim 1 , wherein the model of progression further includes a progression graph.

3. The computer-accessible medium of claim 2 , wherein the progression graph is based on a causal graph.

4. The computer-accessible medium of claim 2 , wherein the model of progression further includes at least one of a directed acyclic graph (DAG), a disconnected DAG, a tree or a forest.

5. The computer-accessible medium of claim 1 , wherein nodes of the model of progression are atomic events and edges that represent a progression between the atomic events.

6. The computer-accessible medium of claim 1 , wherein the model of progression is further based on a noise model.

7. The computer-accessible medium of claim 6 , wherein the noise model includes a biological noise model.

8. The computer-accessible medium of claim 7 , wherein the computer arrangement is further configured to use the biological noise model to distinguish spurious causes from genuine causes.

9. The computer-accessible medium of claim 6 , wherein the noise model includes an experimental noise model.

10. The computer-accessible medium of claim 6 , wherein the noise model includes an experimental noise model and a biological noise model.

11. The computer-accessible medium of claim 1 , wherein the biomedical data further includes imaging data.

12. The computer-accessible medium of claim 1 , wherein the biomedical data includes information pertaining to at least one of at least one normal cell, at least one tumor cell, cell-free circulating DNA or at least one circulating tumor cell.

13. A method for modeling a progression of cancer using biomedical data for one or more patients, comprising:
(a) obtaining the biomedical data which includes at least one of genomics, transcriptomics, or epigenomics;
(b) using a computer hardware arrangement, generating the model of progression which includes (i) states of the cancer and (ii) transitions among the states using a learning network, based on the obtained biomedical data, wherein:
transitions among the states are determined by a causality relationship information whose strength is estimated by probability-raising by at least one estimator which is part of the learning network;
the at least one estimator includes at least one shrinkage estimator, which is a measure of causation among any pair of events atomic events; and
the at least one shrinkage estimator is defined as θ=(1−λ)α(x)+λβ(x), where θ≤λ≤1 can be a shrinkage coefficient, x is an input data, and θ is one or more estimates for an evaluation, wherein the learning network includes the transitions and the at least one shrinkage estimator;

(c) receiving further biomedical data associated with a further patient;
(d) classifying a particular state of cancer for the further patient using the model of progression and the further biomedical data, wherein a particular state of the states of cancer is classified by one or more mutational profiles of the at least one of the genomics, the transcriptomics or the epigenomics; and
(e) determining a genome-specific therapy design based on the classification of the particular state of cancer.

14. A system for modeling a progression of cancer using biomedical data for one or more patients, comprising:
a computer hardware arrangement configured to:
(a) obtain the biomedical data which includes at least one of genomics, transcriptomics, or epigenomics;
(b) using a computer hardware arrangement, generate the model of progression which includes (i) states of the cancer and (ii) transitions among the states using a learning network, based on the obtained biomedical data, wherein:
transitions among the states are determined by a causality relationship whose strength is estimated by probability-raising by at least one estimator which is part of the learning network;
the at least one estimator includes at least one shrinkage estimator, which is a measure of causation among any pair of events atomic events; and
the at least one shrinkage estimator is defined as θ=(1−λ)α(x)+λβ(x), where 0≤λ≤1 can be a shrinkage coefficient, x is an input data, and θ is one or more estimates for an evaluation, wherein the learning network includes the transitions and the at least one shrinkage estimator;

(c) receive further biomedical data associated with a further patient;
(d) classify a particular state of cancer for the further patient using the model of progression and the further biomedical data, wherein a particular state of the states of cancer is classified by one or more mutational profiles of the at least one of the genomics, the transcriptomics or the epigenomics; and
(e) determine a genome-specific therapy design based on the classification of the particular state of cancer.

Assignees

• Univ New York
• Univ Degli Studi Di Milano—Bicocca

Inventors

• Ramazzotti Daniele
• Caravagna Giulio
• Olde Loohuis Loes
• Graudenzi Alex
• Korsuncky Ilya
• Mauri Giancarlo
• Antoniotti Marco
• Mishra Bhubaneswar

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G16H50/50 Primary

  for simulation or modelling of medical disorders · CPC title

• G16B5/00

  ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks · CPC title

• G16B5/20

  Probabilistic models · CPC title

• G06N20/00

  Machine learning · CPC title