Cell-Free DNA Methylation Deconvolution for Tissue-Origin Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing methylation sequencing data from cell-free DNA for cancer detection and monitoring are inadequate, lacking the precision needed for early cancer detection and diagnosis.
Innovation Solution
A deconvolution model is trained using methylation parameters from multiple sources to predict the source of origin of methylation fragments in cell-free DNA, combined with a cancer classifier to determine the likelihood of specific cancer types based on methylation patterns at CpG sites.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If methylation sequencing data is analyzed using conventional methods, then cancer detection can be performed, but the accuracy and precision for early cancer detection and diagnosis are inadequate
Solution Approach 1:
The patent segments the complex methylation sequencing data analysis into distinct components: (1) fragment classification into background, normal, and cancer-associated fragments based on methylation patterns, (2) source deconvolution to identify tissue origins, and (3) cancer type classification. This segmentation enables more precise early cancer detection by systematically processing different aspects of the data separately rather than using conventional holistic methods.
2Loss of information
If deconvolution model with source identification is implemented, then insights into cancer-specific methylation markers can be obtained, but the computational complexity and data processing requirements increase
Solution Approach 1:
The patent applies preliminary action by first classifying fragments into background, normal, and cancer-associated categories before performing source deconvolution. This preliminary classification simplifies the subsequent computational analysis by reducing the complexity of the input data, allowing the deconvolution model to focus specifically on identifying tissue origins of cancer-associated fragments without being overwhelmed by background noise.
Solution Approach 2:
The patent introduces an intermediary deconvolution model that acts as a bridge between raw methylation sequencing data and cancer diagnosis. This intermediary component performs source identification by comparing methylation patterns against reference panels from multiple tissues, thereby recovering source of origin information without requiring direct complex analysis of all raw data.
Data Source
AI summary
A method and system for determining one or more sources of a cell free deoxyribonucleic acid (cfDNA) test sample from a test subject. The cfDNA test sample contains a plurality of deoxyribonucleic acid (DNA) molecules with numerous CpG sites that may be methylated or unmethylated. A trained deconvolution model comprises a plurality of methylation parameters, including a methylation level at each CpG site for each source, and a function relating a sample vector as input and a source of origin prediction as output. The method generates a test sample vector comprising a site methylation metric relating to DNA molecules from the test sample that are methylated at that CpG site. The method inputs the test sample vector into the trained deconvolution model to generate a source of origin prediction indicating a predicted DNA molecule contribution of each source.