Multi-Target Cancer DNA Detection for Low-Frequency Tumor Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting minimal residual disease (MRD) in cancer patients are not sufficiently sensitive due to low levels of tumor DNA in cell-free DNA samples, leading to false positives and challenges in early detection of cancer recurrence.
Innovation Solution
A method that enriches a test sample for multiple target regions with different genetic variations, including single nucleotide variants, multiple nucleotide variants, copy number variants, structural variants, and phased variants, and combines evidence from these regions using error models to enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sequencing-based methods are used to detect tumor DNA in cfDNA, then the approach can identify sequence variations, but the sensitivity is insufficient because the frequency of tumor sequence variations is well below the limit of detection
Solution Approach 1:
The patent segments the detection task by dividing target regions into different classes based on the number and type of genetic variations they contain. This allows tailored error modeling for each class, improving detection sensitivity while controlling false positives through class-specific error probability thresholds.
Solution Approach 2:
The patent applies local quality by assigning different error models to different classes of target regions based on their specific characteristics (number of variants, variant types). Each class receives customized error probability parameters that reflect its specific detection challenges, optimizing both sensitivity and reliability locally.
2Adaptability or versatility
If the number of target regions and genetic variations in an assay is increased, then detection coverage improves, but the potential for false positive results increases
Solution Approach 1:
The patent segments target regions into multiple classes based on the number and types of genetic variations present. This segmentation enables the system to handle diverse variant configurations systematically while applying appropriate error models to each class, preventing false positives from overwhelming the detection system.
Solution Approach 2:
The patent changes the error probability parameters dynamically based on the class of target region being analyzed. By adjusting error thresholds and models according to the specific characteristics of each target region class, the system maintains high reliability even as the number of target regions increases.
3Measurement precision
If sequencing depth is increased to detect low-frequency tumor DNA, then detection sensitivity improves, but the cost and complexity of the assay increases
Solution Approach 1:
The patent applies local quality by determining optimal sequencing depth requirements based on the specific class of target region being analyzed. Different classes may require different sequencing depths based on their error profiles and detection challenges, optimizing the balance between sensitivity and complexity.
Solution Approach 2:
The patent changes sequencing depth parameters adaptively based on the target region class and detection requirements. This allows the system to allocate sequencing resources efficiently, achieving high sensitivity for difficult-to-detect variants without unnecessarily increasing depth for easier detections.
Data Source
AI summary
In one embodiment, the method may comprise enriching the test sample for a plurality of target regions, wherein the plurality of target regions comprises a first target region having a first class and a second target region having a second class. The plurality of target regions may be measured and for each of the first target region and second target region, the measurements that support the class of the target region may be compared to an error model that models the probability of observing that class of target region in DNA that does not contain that class of target region. These comparisons may then be combined for at least the first target region and the second target region. Cancer DNA may then be identified in the test sample based on the combined comparisons.