HRD Detection via Mutational Spectrum and ML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting homologous recombination deficiency (HRD) in tumor tissues are limited in accuracy and predictive value, as negative results do not guarantee lack of response to PARP inhibitors, and there is a lack of consensus on defining and measuring HRD components, making it difficult to determine which tumors would benefit from PARP inhibitors or platinum-based chemotherapy.
Innovation Solution
A method using trained machine learning classifiers to identify HRD from omics data by generating a mutational spectrum from tumor samples, employing K-means clustering to group optimal clusters, and determining HRD scores, which can indicate responsiveness to PARP inhibitors or platinum-based chemotherapy, even in tumors without BRCA1/2 mutations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current HRD detection methods are used, then detection can be performed, but accuracy and predictive value are limited
Solution Approach 1:
The patent segments the HRD detection process into multiple independent components: LOH (loss of heterozygosity) analysis, TAI (telomeric allelic imbalance) analysis, and LST (large-scale state transitions) analysis. Each component is measured separately using specific genomic scar metrics, and then combined to generate a comprehensive HRD score. This segmentation allows for more precise measurement of each aspect of HRD while maintaining overall reliability through the composite scoring system.
2Ease of manufacture
If traditional HRD assay methods are used, then results can be obtained, but there is lack of consensus on definition and measurement of components
Solution Approach 1:
The patent establishes specific parameter thresholds and measurement criteria for each HRD component. LOH is measured using allele frequency deviations with defined cutoff values, TAI is quantified using telomeric region copy number variations with specific thresholds, and LST is identified through breakpoint analysis with defined criteria. These standardized parameter changes enable consistent measurement across different laboratories and assays, resolving the lack of consensus while maintaining measurement precision.
3Measurement precision
If machine learning with signature multivariate analysis is used, then HRD detection is improved, but it is limited to BRCA1/2 mutations
Solution Approach 1:
The patent creates a universal HRD detection system that functions across multiple cancer types and genetic backgrounds. By measuring genomic scars (LOH, TAI, LST) that result from homologous recombination deficiency regardless of its cause, the assay can detect HRD in BRCA1/2 mutated tumors, BRCA wild-type tumors with somatic HRD pathway mutations, and even tumors with epigenetic silencing of HR genes. This multi-functional approach maintains high measurement precision while achieving broad adaptability across different tumor types and genetic etiologies.
Data Source
AI summary
Disclosed herein are methods of identifying homologous recombination deficiency (HRD) in omics data, comprising generating a mutational spectrum from omics data; and using the mutational spectrum in a trained model to identify HRD. Further disclosed herein are methods of treating a tumor that has HRD score indicating significant HRD events, comprising: obtaining omics data from a tumor sample and generating a mutational spectrum from omics data; using the mutational spectrum in a trained model to identify HRD in the omics data from the tumor sample; identifying the cancer as likely responsive to treatment with a PARP inhibitor upon determination of HRD; and administering a PARP inhibitor treatment for the tumor upon determination of a high HRD score.