Interrogatory cell-based assays and uses thereof
Inventors
Narain, Niven Rajin • Sarangarajan, Rangaprasad • VISHNUDAS, VIVEK K.
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11456054-B2
Publication Date
2022-09-27
Expiration Date
Abstract
Described herein is a discovery Platform Technology for analyzing a biological system or process (e.g., a disease condition, such as cancer) via model building.
Core Innovation
The invention provides a method for identifying a modulator of a disease process by building a computer-implemented causal relationship network model from multi-omics data and measured functional activity or measured cellular response of disease-related cells. The method obtains measured expression levels of one or more genes, measured lipidomics data, measured metabolomics data, or a combination, together with a second data set representing a measured functional activity or a measured cellular response.
A first causal relationship network model is generated in a programmed computing system using network fragment scoring and an ensemble of trial networks. The causal relationships and causality are determined based on the first data set and the second data set and not based on previously identified or known biological relationships between variables, by creating network fragments with variables connected by one or more relationships, determining probabilistic scores, and evolving the trial networks in parallel through global optimization.
From the computer-implemented causal relationship network model, the method identifies a causal relationship unique in the disease process, and a gene, a lipid, or a metabolite associated with the unique causal relationship is identified as a modulator of a disease process. The disclosed platform supports generating disease-unique causal relationships from multi-omics cellular model data, and includes Bayesian scoring, Bayesian network representation, in silico simulation, differential causal relationship networks, and disease-environment conditions including hypoxia, hyperglycemia, and lactic acid-rich culture condition.
Claims Coverage
The consolidated claim coverage includes one independent claim for identifying a modulator of a disease process and dependent refinements. The inventive features center on multi-omics and functional response data, unbiased causal relationship network generation, network fragment scoring with trial networks, and identification of disease-unique causal relationships as modulators.
Disease modulator identification from multi-omics and functional response
Obtain a first data set representing measured expression levels of one or more genes, measured lipidomics data, measured metabolomics data, or a combination, and obtain a second data set representing a measured functional activity or a measured cellular response.
Unbiased causal relationship network generation without known biological relationships
Generate a computer-implemented first causal relationship network model relating gene expression, lipidomics or metabolomics data, and functional activity or cellular response based on the first data set and the second data set, wherein relationships and causality are determined based on the data and not based on previously identified or known biological relationships between variables.
Network fragment scoring and evolved ensemble of trial networks
Create a list of network fragments and determine a probabilistic score for each network fragment based on the first data set and/or the second data set; create an ensemble of trial networks using different subsets of the network fragments; globally optimize the ensemble by evolving the trial networks in parallel using multiple processors.
Identification of disease-unique causal relationships as modulators
Identify, from the causal relationship network model, a causal relationship unique in the disease process, and identify a gene, a lipid, or a metabolite associated with the unique causal relationship as a modulator of a disease process.
Bayesian scoring and Bayesian trial networks
The probabilistic score is a Bayesian score and each trial network is a Bayesian network.
In silico simulation refinement with prediction confidence level
Refine the consensus relationship network model through in silico simulation to generate a simulated causal relationship network model with a prediction confidence level for one or more causal relationships.
Disease-environment conditions for measuring disease-related cells
The method is performed in an environment that includes hypoxia, hyperglycemia, lactic acid-rich culture condition, or combinations thereof.
Differential causal relationship network construction
Generate a computer-implemented differential causal relationship network by comparing relationships between the same pairs of nodes across two causal relationship network models, marking relationships absent from the other model and relationships with at least one significantly different parameter.
Energy metabolism pathway shift from glycolysis to oxidative phosphorylation
The modulator shifts energy metabolism in disease cells from glycolysis toward oxidative phosphorylation.
Coverage centers on constructing an unbiased, data-driven causal relationship network model from multi-omics measurements and functional activity or cellular response to find a disease-unique causal relationship, then identifying a corresponding gene, lipid, or metabolite as a modulator. Dependent refinements specify Bayesian and Bayesian-network implementations, in silico simulation with prediction confidence, disease-environment perturbation contexts, differential network comparison logic, and an energy metabolism pathway shift association.
Stated Advantages
Enables identification of a causal relationship unique in the disease process.
Enables identification of a gene, a lipid, or a metabolite associated with a disease-unique causal relationship as a modulator of a disease process.
Determines relationships and causality based on obtained data rather than previously identified or known biological relationships.
Supports constructing differential causal relationship networks by comparison across two causal relationship network models.
Supports generating a simulated causal relationship network model with a prediction confidence level.
Supports disease-context evaluation using environmental perturbations such as hypoxia, hyperglycemia, and lactic acid-rich culture conditions.
Documented Applications
Identifying disease-unique causal relationships and associated modulators or drivers for therapeutic targets.
Identifying biomarkers.
Profiling drug efficacy.
Profiling drug toxicity, including drug-induced cardiotoxicity risk and mitigation.
Constructing and comparing differential causal networks between disease and normal or treated and untreated conditions to support interpretation of disease processes.
Interested in licensing this patent?