Machine Learning Model for Genetic Variant Quality Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques struggle to identify and select high-quality somatic genetic variants that will sufficiently amplify in tumour-informed assays, leading to a substantial proportion of poor-quality variants being included.
Innovation Solution
A computational process using a trained machine learning model to assess the quality of genetic variants based on a plurality of characteristics, determining their likelihood of being somatic and amplifying sufficiently for detection in a ctDNA assay.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If existing techniques are used to identify genetic variants, then a large number of variants can be identified, but a substantial proportion (e.g., 1/3) of the variants are poor quality and will not be detected in patient blood samples
Solution Approach 1:
The patent applies preliminary action by training a machine learning model in advance using characteristics from genetic variants that were previously validated through amplicon sequencing and digital PCR. The model learns to predict which variants will successfully amplify and be detected, allowing quality assessment to be performed before the tumour-informed assay is executed, thus preventing poor-quality variants from being included in the final panel
Solution Approach 2:
The patent replaces the mechanical/experimental validation process (amplicon sequencing and digital PCR testing of each variant) with a computational machine learning model. Instead of physically testing each genetic variant through complex wet-lab procedures, the system uses a trained ML model that processes variant characteristics computationally to predict amplification success and detection reliability
2Measurement precision
If more genetic variants are included in the panel to improve sensitivity, then the assay can detect more potential MRD indicators, but the proportion of poor-quality variants increases leading to false negatives
Solution Approach 1:
The patent implements feedback by using the results of amplicon sequencing and digital PCR validation to train the machine learning model. The model learns from historical data about which variants succeeded or failed in detection, and this learned knowledge is fed back into the variant selection process to improve future predictions of variant quality and amplification success
3Productivity
If traditional filtering methods are used to select genetic variants, then the selection process is simple and fast, but the ability to accurately predict amplification success and detection quality is insufficient
Solution Approach 1:
The patent replaces traditional mechanical filtering methods with a machine learning-based computational system. The ML model processes multiple variant characteristics simultaneously and predicts quality outcomes with high accuracy, maintaining computational efficiency while dramatically improving prediction accuracy compared to simple threshold-based filtering
Data Source
AI summary
Methods and apparatus for assessing quality of a genetic variant for inclusion as a biomarker in a circulating tumour DNA (ctDNA) assay. The method includes receiving genetic variants associated with a sample collected from a patient, the set of genetic variants including a first genetic variant, determining values for a plurality of characteristics associated with the first genetic variant, providing the values for the plurality of characteristics as input to a trained machine learning (ML) model, the trained ML model being trained to output a quality of a genetic variant, the quality of the genetic variant representing a likelihood that the genetic variant is both somatic and will sufficiently amplify using amplicon sequencing, and including the first genetic variant in a panel of genetic variants for use in a ctDNA assay for the patient based on the quality of the first genetic variant output from the trained ML model.