State-Based Base Calling Using Prior Cycle Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deploying deep convolution neural networks (CNNs) on portable and embedded systems is challenging due to large data volume, intensive computation, varying algorithm structures, and frequent memory accesses, which affects the efficiency and performance of hardware accelerators like FPGAs, and existing base callers suffer from k-mer biases and memory limitations during base calling, reducing accuracy.
Innovation Solution
Incorporate state information from previous sequencing cycles into the base calling process using state-based methods, generating various types of states through mathematical manipulations and algorithms to enhance base calling accuracy and reduce memory requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep convolution neural networks are deployed on portable and embedded systems, then base calling accuracy can be improved, but device complexity and memory requirements increase significantly
Solution Approach 1:
The patent segments the base calling process into multiple passes, where each pass processes a subset of k-mers. This divides the complex task of analyzing all possible k-mers in a sequence into manageable chunks, reducing the computational burden on hardware accelerators while maintaining accuracy. The segmentation allows the system to process data in smaller batches rather than attempting to handle the entire sequence at once.
Solution Approach 2:
The patent performs preliminary actions by pre-calculating and storing k-mer frequency counts and other features before the main base calling process. This preliminary processing reduces the amount of computation required during actual base calling, as the system can leverage pre-computed statistics rather than performing all calculations in real-time, thereby reducing device complexity requirements.
2Measurement precision
If deep convolution neural networks are deployed on portable and embedded systems, then base calling accuracy can be improved, but memory requirements increase
Solution Approach 1:
The patent extracts only the essential features and statistics needed for accurate base calling, discarding redundant information. By selectively extracting and storing only the most important k-mer frequency counts and signal characteristics rather than retaining all raw data and intermediate computations, the system significantly reduces memory requirements while maintaining the accuracy needed for reliable base calling.
Solution Approach 2:
The patent performs preliminary computation of statistical features and stores only the compressed results rather than retaining all原始 data. This preliminary processing step creates a compact representation of the sequencing data that maintains informational content while dramatically reducing memory footprint, allowing accurate base calling with reduced memory requirements.
3Productivity
If existing base callers are used, then processing speed can be maintained, but k-mer biases reduce base calling accuracy
Solution Approach 1:
The patent incorporates feedback mechanisms that continuously monitor and adjust base calling decisions based on observed k-mer frequencies and signal patterns. By feeding back information about regional biases and adjusting the base calling algorithm in response to these patterns, the system can compensate for k-mer biases while maintaining processing speed. The feedback loop allows real-time correction of systematic errors without requiring complete reprocessing.
Data Source
AI summary
The technology disclosed relates to state-based base calling. In particular, the technology disclosed relates to incorporating state information about data from previous sequencing cycles into the analysis of data from a current sequencing cycle when generating a base call for the current sequencing cycle. For example, when generating a base call for an Nth sequencing cycle, the technology disclosed can incorporate into the base calling logic state information about data from sequencing cycles 1 to N−1.