Sequence Pattern Detection for False Variant Call Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Next-generation sequencing data is highly correlated and has complex interdependencies, hindering the application of traditional classifiers like support vector machines for accurate variant filtering, leading to sequence-specific errors (SSEs) that result in false variant calls and assembly gaps.
Innovation Solution
A deep learning-based framework, DeepPOLY, utilizing a hierarchical structure of convolutional and fully-connected neural networks to identify sequence patterns causing SSEs by training on large-scale variant data, simulating known patterns, and analyzing parameter activations to detect SSE-causing sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional classifiers like support vector machines are used for variant filtering, then the classification process is simple and fast, but the accuracy is insufficient due to high correlation and complex interdependencies in sequencing data
Solution Approach 1:
The patent replaces traditional mechanical/statistical classification systems (support vector machines) with a neural network-based system that can automatically capture complex patterns in sequencing data. The neural network learns intricate relationships between sequence features and variant calls without requiring manual feature engineering, thereby improving accuracy while managing complexity through automated learning.
Solution Approach 2:
The patent transforms the classification approach by changing from fixed-parameter traditional classifiers to adaptive-parameter neural networks. The neural network dynamically adjusts its internal parameters (weights and biases) during training to optimize for the specific characteristics of sequencing data, enabling it to handle high correlation and complex interdependencies that static classifiers cannot capture.
2Measurement precision
If deep learning-based framework is used to identify sequence patterns, then the accuracy in identifying SSE-causing patterns is improved, but the computational complexity and training time increase
Solution Approach 1:
The patent performs preliminary action by pre-training the neural network on large-scale sequencing data to learn general sequence patterns and error characteristics. This pre-training phase enables the model to capture fundamental relationships before being fine-tuned on specific tasks, reducing the time required for subsequent specialized training while maintaining high accuracy in identifying SSE-causing patterns.
3Reliability
If deep neural networks are used for variant filtering, then the ability to handle complex interdependencies is improved, but the computational resources and processing power required increase
Solution Approach 1:
The patent segments the variant filtering process into distinct neural network components that can process different aspects of the sequencing data independently. By dividing the complex task into manageable segments (e.g., feature extraction, pattern recognition, classification stages), the system can process complex interdependencies more efficiently with reduced computational overhead compared to a monolithic approach.
Data Source
AI summary
The technology disclosed presents a deep learning-based framework, which identifies sequence patterns that cause sequence-specific errors (SSEs). Systems and methods train a variant filter on large-scale variant data to learn causal dependencies between sequence patterns and false variant calls. The variant filter has a hierarchical structure built on deep neural networks such as convolutional neural networks and fully-connected neural networks. Systems and methods implement a simulation that uses the variant filter to test known sequence patterns for their effect on variant filtering. The premise of the simulation is as follows: when a pair of a repeat pattern under test and a called variant is fed to the variant filter as part of a simulated input sequence and the variant filter classifies the called variant as a false variant call, then the repeat pattern is considered to have caused the false variant call and identified as SSE-causing.