Fetal DNA Analysis via Random Forest Regression

Solution Overview

Problem

Current methods for determining fetal DNA in maternal plasma samples using massive sequencing are limited by low precision, high error rates, and a high number of false positives, particularly in assessing chromosomal and chromosome fragment alterations and the presence or absence of the Y chromosome.

Innovation Solution

A method utilizing a Random Forest prediction model that incorporates gestational week, mother's height, and mother's weight to identify genomic regions enriched in fetal DNA, reducing errors and improving sensitivity by focusing on nucleosome profiles and adjusting for physiological variables.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional methods are used to determine fetal DNA percentage in maternal plasma, then the analysis can be performed, but the precision is low and error rates are high

Engineering Contradiction:

Improvefetal DNA percentage determination precisionVSAvoiderror rate in chromosomal alteration assessment

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The genome is divided into multiple bins of varying sizes (e.g., 50-200 Kb regions), allowing localized analysis of sequence read coverage across different genomic regions. This segmentation enables more precise detection of chromosomal alterations by examining coverage patterns in specific regions rather than treating the entire genome as a single unit.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces multiple dimensions to the analysis by incorporating maternal physiological parameters (height, weight, gestational age) and environmental factors (heparin intake) as co-variables in the regression model. This multi-dimensional approach allows the system to account for confounding factors that affect fetal DNA fraction, thereby improving measurement precision and reducing false positives.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Reliability

If conventional sequencing analysis is used, then chromosomal alterations can be detected, but the number of false positives is high

Engineering Contradiction:

Improvedetection accuracy of chromosomal alterationsVSAvoidfalse positive rate in Y chromosome detection

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent employs regression models that use maternal physiological parameters and environmental factors as feedback variables to adjust and refine the estimation of fetal DNA fraction. This feedback mechanism allows the system to continuously improve its accuracy by accounting for factors that influence fetal DNA characteristics, thereby reducing false positives in chromosomal alteration detection.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The analysis dynamically adjusts parameters such as bin sizes, coverage thresholds, and statistical significance levels based on the specific sample characteristics and maternal parameters. This adaptive parameter adjustment optimizes the detection sensitivity and specificity for each individual case, reducing false positives while maintaining high detection accuracy.

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If comprehensive genomic analysis is performed, then accurate fetal DNA assessment is achieved, but the computational complexity increases

Engineering Contradiction:

Improvefetal DNA fraction determination accuracyVSAvoidbioinformatics analysis complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The genome is divided into multiple bins of varying sizes (e.g., 50-200 Kb regions), allowing localized analysis of sequence read coverage across different genomic regions. This segmentation enables more precise detection of chromosomal alterations by examining coverage patterns in specific regions rather than treating the entire genome as a single unit.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces multiple dimensions to the analysis by incorporating maternal physiological parameters (height, weight, gestational age) and environmental factors (heparin intake) as co-variables in the regression model. This multi-dimensional approach allows the system to account for confounding factors that affect fetal DNA fraction, thereby improving measurement precision and reducing false positives.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3884502B1Method and computer program product for analysis of fetal DNA by massive sequencing

Publication Date: 2022.11.09 SISTEMAS GENOMICOS

Figure 1~2
Figure 3
Figure 3

AI summary

Method of analysis of fetal DNA by massive sequencing and computer program product to implement said method of analysis. The method is for determining the percentage of fetal DNA in a sample, for determining whether a fetus having fetal DNA is at risk of having alterations in the number of at least one chromosome and/or at least one chromosome fragment, wherein said alterations in the number of chromosomes are complete chromosomal fetal aneuploidies and said alterations in the number of chromosome fragments are partial fetal aneuploidies; and for determining the absence or presence of Y chromosome.