Machine-Learning HRD Classification From Tumor and Normal DNA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods lack effective computational resources to predict homologous recombination deficiency (HRD) in cancer patients, limiting the identification of those likely to respond favorably to PARP inhibitors and other targeted therapies.
Innovation Solution
A machine-learning algorithm is trained using DNA sequencing data from cancerous and non-cancerous tissues to predict HRD status by analyzing features such as heterozygosity status, loss of heterozygosity, and variant alleles in DNA damage repair genes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional treatment methods are used for all patients with the same cancer type, then treatment simplicity is maintained, but treatment effectiveness and patient outcomes deteriorate
Solution Approach 1:
The patent performs genomic sequencing and HRD status prediction before treatment selection, enabling physicians to identify suitable patients for PARP inhibitors in advance. This preliminary genomic characterization allows treatment personalization without adding complexity to the actual treatment administration process.
Solution Approach 2:
The patent replaces complex clinical judgment and trial-and-error treatment selection with an automated machine learning classifier that processes genomic data and predicts HRD status. This computational approach substitutes subjective medical decision-making with an objective, reproducible algorithmic system.
2Reliability
If genomic testing is performed for all patients to enable precision oncology, then treatment personalization improves, but cost and resource requirements increase
Solution Approach 1:
The patent extracts and analyzes only the specific genomic features relevant to HRD status (loss of heterozygosity, telomeric allelic imbalance, and heterozygosity status) from the complete genomic sequence. This selective extraction of critical information reduces computational burden and resource requirements while maintaining prediction accuracy.
Solution Approach 2:
The patent transforms complex genomic sequence data into simplified binary classifications (HRD-positive or HRD-negative) based on threshold criteria. This parameter transformation converts continuous genomic measurements into discrete, actionable clinical categories that require minimal computational resources for interpretation.
3Measurement precision
If comprehensive genomic analysis is performed to identify HRD status, then prediction accuracy improves, but computational complexity and time requirements increase
Solution Approach 1:
The patent divides the genomic analysis into three independent feature categories: loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and heterozygosity status (HS). Each feature is calculated separately using dedicated algorithms, allowing parallel processing and reducing overall computation time compared to a monolithic analysis approach.
Solution Approach 2:
The patent implements a tiered analysis approach where the machine learning classifier can function with varying levels of genomic data completeness. The system can produce HRD predictions using subsets of the full genomic feature set, enabling rapid preliminary assessments when data is limited while maintaining the option for comprehensive analysis when resources permit.
Data Source
AI summary
Methods, systems, and software are provided for determining a homologous recombination pathway status of a cancer in a test subject, e.g., to improve cancer treatment predictions and outcomes. In some embodiments, classifiers using one or more of (i) a heterozygosity status for DNA damage repair genes in a cancerous tissue, (ii) a measure of the loss of heterozygosity across the genome of the cancerous tissue, (iii) a measure of variant alleles detected in a second plurality of DNA damage repair genes in the genome of the cancerous tissue, (iv) a measure of variant alleles detected in the second plurality of DNA damage repair genes in the genome of a non-cancerous tissue, and (v) tumor sample purity are provided.