Two-Step Variant Caller Using Cell-Specific Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional variant calling methods suffer from high false positive rates due to sequencing and PCR errors, leading to the elimination of true positives, necessitating improved techniques for error correction and variant identification.
Innovation Solution
A two-step process involving a cell-specific error correction model to correct sequence reads and a variant caller model to analyze corrected reads, enhancing the accuracy of variant calls by distinguishing true from false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hard cutoffs are used to mitigate false positives, then false positive rate is reduced, but true positive rate decreases (missing data problem)
Solution Approach 1:
The patent applies preliminary error correction to sequence reads before variant calling. By correcting sequencing and PCR errors in advance through a two-step process (error correction model followed by variant calling model), the system prepares cleaned data that reduces false positives without sacrificing true positives, resolving the contradiction between reliability and measurement precision
Solution Approach 2:
The patent introduces an intermediary error correction model between the raw sequencing data and the variant calling process. This intermediary component processes sequence reads through cell-specific error correction using machine learning models, acting as a mediator that cleans the data before it reaches the variant caller, thereby reducing false positives while maintaining true positive detection
2Productivity
If conventional variant calling methods are used, then processing speed is maintained, but accuracy of variant identification deteriorates due to high false positive rates
Solution Approach 1:
The patent performs error correction as a preliminary action before variant calling. The two-step process (error correction model then variant calling model) is designed to be computationally efficient while significantly improving accuracy. By preparing corrected sequence reads in advance, the system maintains processing speed through optimized workflows while achieving higher variant identification accuracy
Solution Approach 2:
The patent replaces conventional mechanical filtering approaches with machine learning-based error correction models. Instead of using simple hard cutoffs or traditional statistical filters, the system employs trained neural network models that can accurately distinguish between true variants and errors, thereby improving accuracy while maintaining computational efficiency through modern ML optimizations
Data Source
AI summary
Described herein are improved variant calling methods including a two-step process involving 1) error correction of bases in sequence reads through a cell-specific process and 2) variant calling across cell populations using the error corrected sequence reads. Generally, the first step of error correction involves applying a first machine learned model to identify and correct bases of sequence reads. The second step of variant calling involves applying a second machine learned model to classify a base. Such improved variant calling methods can be useful for identifying variants that are implicated in biological processes, such as diseased biological processes.