Epigenetic Analysis via Parametric Statistical Model

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for epigenetic analysis fail to effectively account for stochasticity and uncertainty in DNA methylation, particularly in the non-independent behavior among methylation sites, leading to limited resolution and accuracy in understanding epigenetic changes and their role in health and disease.

Innovation Solution

A method is developed to calculate the epigenetic potential energy landscape (PEL) by partitioning the genome into discrete regions, analyzing methylation status using a parametric statistical model that accounts for dependence among methylation sites, and computing the corresponding joint probability distribution, enabling detailed epigenetic analysis, including classification and identification of genomic features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If empirical methods are used to measure entropy in DNA methylation, then measurement can be performed, but resolution is limited and extensive cell culture expansion and molecular tagging are required

Engineering Contradiction:
Improveentropy measurement resolutionVSAvoidcell culture expansion and molecular tagging requirements
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces complex wet-lab procedures (cell culture expansion and molecular tagging) with computational methods. Specifically, it uses probabilistic models and entropy calculations on existing methylation data to achieve high-resolution entropy measurement without requiring extensive physical manipulation of biological samples

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces computational intermediaries (probabilistic models, entropy calculations, and bioinformatics pipelines) that bridge the gap between raw methylation data and meaningful epigenetic insights, eliminating the need for direct physical intervention through cell culture and molecular tagging

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If simple probabilistic models are used that model methylation sites independently, then computational simplicity is achieved, but the non-independent behavior among methylation sites is not accounted for

Engineering Contradiction:
Improvemodel complexityVSAvoidaccuracy in capturing methylation site dependence
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent segments the genome into discrete regions and applies probabilistic models at the regional level rather than treating all methylation sites independently. This allows capturing local dependencies among nearby sites while maintaining computational tractability through regional decomposition

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent employs dynamic probabilistic models that can adapt to local variations in methylation patterns. By allowing model parameters to vary across different genomic regions and incorporating spatial dependencies, the model dynamically adjusts to capture non-independent behavior without requiring excessive complexity

Inventive Principle:
Principle #15Dynamics

3Reliability

If deterministic models with noise terms are used to account for epigenetic uncertainty, then stochasticity is incorporated, but the models remain relatively simple and do not fully capture complex epigenetic patterns

Engineering Contradiction:
Improveaccounting for stochasticity and uncertaintyVSAvoidmodel sophistication
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent transforms the modeling approach by changing from simple noise terms to comprehensive probabilistic parameters that characterize entire methylation patterns. By using parameters like entropy, probability distributions, and regional methylation states, the model captures complex epigenetic patterns while maintaining mathematical tractability

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent creates a composite modeling framework that integrates multiple components: probabilistic models, entropy calculations, regional segmentation, and spatial dependency structures. This composite approach achieves high reliability in capturing stochasticity and complex patterns without requiring any single component to be excessively complex

Inventive Principle:
Principle #40Composite materials

Data Source

PatentEP3472358B1Methods and system for epigenetic analysis
Publication Date: 2021.08.18 JOHNS HOPKINS UNIVERSITY
  • EP3472358B1 patent drawingFigure 1A~1B
  • EP3472358B1 patent drawingFigure 1C
  • EP3472358B1 patent drawingFigure 2A

AI summary

The present disclosure provides computational methods for epigenetic analysis as well as systems for implementing such analyses.