Chromatin Signal V-Plot Prediction for Cell-Free DNA

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for revealing biological attributes using nucleic acid sequence signals are limited in their ability to accurately predict features such as phenotypes, disorders, and tissue origins from cell-free nucleic acids.

Innovation Solution

The method involves deriving chromatin-related nucleic acid signals from cell-free DNA or RNA, constructing sequence signature V-plots, and training machine-learning models to predict biological attributes, including neoplasia, inflammation, tissue damage, and neurodegeneration.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional nucleic acid sequencing methods are used to predict biological attributes from cell-free nucleic acids, then the process is simple and quick, but the prediction accuracy is insufficient

Engineering Contradiction:
Improveprediction accuracyVSAvoidmethod complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the nucleic acid sequence into regulatory regions (promoters, enhancers, silencers, etc.) and analyzes each region's chromatin signals separately. This segmentation allows the system to capture specific biological attributes from different genomic locations, improving prediction accuracy without requiring analysis of the entire genome at once.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from traditional 1D sequence analysis to 2D V-plot visualization and 3D frequency signal map representation. By adding dimensional layers (fragment length vs. position, frequency signals across multiple dimensions), the system reveals patterns in chromatin-related factors that are not visible in conventional linear sequence views, thereby improving prediction accuracy.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If deep learning frameworks are applied to analyze chromatin signals, then prediction accuracy improves, but computational resources and processing time increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidcomputational energy
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent performs preliminary processing by extracting and pre-aligning chromatin-related nucleic acid signals to regulatory regions before feeding data into the deep learning model. This pre-processing step organizes the data in advance, reducing the computational burden during model training and inference while maintaining high prediction accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates simplified representations (V-plots and frequency signal maps) that copy the essential patterns from raw sequencing data into a compressed format. These visual representations serve as efficient proxies that capture biological attributes with less computational complexity than processing the full raw data through deep learning models.

Inventive Principle:
Principle #26Copying

3Measurement precision

If chromatin-related factors are analyzed in detail, then biological attribute prediction improves, but the complexity of data processing increases

Engineering Contradiction:
Improveprediction accuracyVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces V-plots and frequency signal maps as intermediary representations that mediate between raw sequencing data and biological attribute predictions. These intermediaries simplify the complex chromatin signal data into visually interpretable patterns, making the data processing pipeline more manageable while preserving the essential information needed for accurate predictions.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3765017B1Methods using nucleic acid signals for revealing biological attributes
Publication Date: 2025.05.21 THE BOARD OF TRUSTEES OF THE LELAND STANFORD JUNIOR UNIV
  • EP3765017B1 patent drawingFigure 1
  • EP3765017B1 patent drawingFigure 2
  • EP3765017B1 patent drawingFigure 3

AI summary

Processes to reveal biological attributes from nucleic acids are provided. In some instances, nucleic acids are used to develop frequency sequence signal maps, construct V-plots, and/or to train computational models. In some instances, trained computational models are used to predict features that reveal biological attributes.