Gene Mutation Detection Model for Low-Frequency Variant Calling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gene mutation detection methods struggle to achieve both high precision and high recall, particularly for mutation frequencies less than 1.5%, due to sequencing biases and batch variations, leading to inaccurate detection results.
Innovation Solution
A method involving multiple mutation detection modules and a pre-trained target mutation detection model, where suspected mutation sites are identified with high recall by a first module, further refined by a second module for high precision, and integrated with sequencing data processing to ensure accurate mutation detection results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional single-module mutation detection methods are used, then the detection process is simple, but both precision and recall cannot be ensured simultaneously, especially for low-frequency mutations
Solution Approach 1:
The detection system is divided into multiple specialized modules: a first mutation detection module that prioritizes recall to identify all potential mutation sites, a second mutation detection module that prioritizes precision to filter false positives, and a target mutation detection model that integrates results from both modules. This segmentation allows each module to optimize for its specific function rather than trying to balance both precision and recall in a single system.
Solution Approach 2:
The system merges the outputs of the first and second mutation detection modules through the target mutation detection model. The target model integrates mutation feature data from both modules, combining the comprehensive site identification capability of the first module with the filtering capability of the second module to achieve both high precision and high recall simultaneously.
2Measurement precision
If multiple mutation detection modules are combined to improve precision and recall, then detection accuracy improves, but the system complexity increases
Solution Approach 1:
The target mutation detection model serves as an intermediary that receives and integrates results from the first and second mutation detection modules. It processes mutation feature data from both modules and produces the final detection results, managing the complexity of combining multiple modules while achieving improved precision and recall.
3Reliability
If the first mutation detection module prioritizes recall to identify all suspected mutation sites, then fewer true mutations are missed, but more false positive sites are identified
Solution Approach 1:
The system segments the detection task into two phases: the first module focuses exclusively on maximizing recall to identify all potential mutation sites without worrying about false positives, while the second module and target model handle the precision aspect by filtering and validating the suspected sites identified in the first phase.
4Measurement precision
If the second mutation detection module performs feature extraction to improve precision, then false positives are reduced, but some true mutations may be missed
Solution Approach 1:
The target mutation detection model merges the results from both modules, combining the comprehensive site list from the first module (high recall) with the filtered results from the second module (high precision). This integration ensures that true mutations identified by the first module are not lost, while false positives are reduced through the combined analysis.
Data Source
AI summary
Provided are a gene mutation detection method and apparatus, a device, a medium, and a product. The method includes acquiring a suspected mutation site of a nucleic acid sample under test, where the suspected mutation site is determined based on first mutation feature data generated by a first mutation detection module upon mutation calling performed on sequencing data of the nucleic acid sample under test, and the recall at which the first mutation detection module identifies gene mutation sites is greater than or equal to a preset recall; acquiring second mutation feature data and third mutation feature data of each suspected mutation site; and inputting the second mutation feature data and the third mutation feature data into a pre-trained target mutation detection model and outputting a mutation detection result of each suspected mutation site.