Somatic Mutation Clonality Estimation with Variational Bayesian Mixtures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating cancer cell fractions (CCF) are inefficient, inaccurate, and computationally demanding due to assumptions about clonality and copy number variations (CNVs), particularly when dealing with subclonal mutations and mixed tumor samples.
Innovation Solution
The implementation of variational Bayesian mixture models to cluster initial cancer cell fractions (iCCFs) and determine final cancer cell fractions (fCCFs) while accounting for tumor purity and copy number variations, using a method that does not rely on Markov chain Monte Carlo (MCMC) techniques, allowing for faster and more accurate estimation of CCFs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Markov chain Monte Carlo (MCMC) techniques are used to estimate cancer cell fractions, then measurement precision can be improved, but computational time and device complexity increase significantly
Solution Approach 1:
The patent replaces the complex Markov chain Monte Carlo (MCMC) computational system with a variational Bayesian mixture model system. This substitution maintains the ability to estimate cancer cell fractions accurately while dramatically reducing computational time and complexity, as the variational approach provides closed-form solutions rather than requiring iterative sampling processes.
Solution Approach 2:
The patent changes the mathematical parameters and computational approach by using variational Bayesian inference with mixture models. This parameter change transforms the estimation problem from one requiring complex MCMC simulations to one that can be solved more efficiently using variational methods, achieving both accuracy and speed.
2Productivity
If simple nucleotide variant calling methods are used, then productivity is improved, but measurement precision deteriorates due to inability to handle copy number variations
Solution Approach 1:
The patent creates a universal variant calling system that can handle multiple types of genetic variations including simple nucleotide variants and copy number variations within a single integrated framework. The variational Bayesian mixture model simultaneously accounts for different mutation types and their implications for cancer cell fraction estimation, eliminating the need for separate analysis pipelines.
Solution Approach 2:
The patent combines multiple data types and analytical approaches into a composite system that integrates copy number variation information with variant allele frequency data. This composite approach allows the system to maintain high productivity while improving measurement precision by considering both simple variants and CNVs in a unified analysis.
3Device complexity
If clonality assumptions are made to simplify analysis, then device complexity is reduced, but reliability deteriorates when dealing with subclonal mutations
Solution Approach 1:
The patent introduces dynamic modeling of clonality through variational Bayesian mixture models that can adapt to different clonal structures in the data. Rather than assuming a fixed clonal architecture, the model dynamically infers the number and characteristics of clones present in the tumor sample, improving reliability while maintaining manageable complexity through variational approximation.
Solution Approach 2:
The patent implements feedback mechanisms where the estimated cancer cell fractions from initial variant calls are used to refine subsequent analysis. The variational Bayesian model iteratively updates its estimates based on observed data patterns, allowing the system to correct for subclonal structures and improve overall reliability without requiring overly complex initial assumptions.
Data Source
AI summary
Computer implemented methods and computer systems are provided for estimating cancer cell fractions indicating proportions of cancer cells carrying one or more mutations of interest using one or more nucleic acid samples from a subject. The methods and systems provided herein implement processes that use a variational Bayesian mixture model to cluster initial cancer cell fractions and obtain the one or more final cancer cell fractions, the initial cancer cell fractions accounting for cancer purity and copy numbers. The disclosed methods and systems improve accuracy, validity, and reliability of tests for cancer clonality, and save time, materials, cost, and computer resources required for the tests, which can help design more affective cancer treatments.