Base Calling Using Multiple Base Caller Models on FPGA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deploying deep Convolutional 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 of Graphics Processing Unit (GPU) and other general-purpose platforms, necessitating specialized acceleration hardware like Field Programmable Gate Arrays (FPGAs) for efficient convolution operations.
Innovation Solution
A system utilizing multiple base callers, including neural network-based and non-neural network-based models, is implemented on a configurable processor like an FPGA to accelerate base calling operations in sequencing systems, leveraging customized dataflow and hardware architecture to optimize resource utilization and minimize data communication, thereby enhancing the performance of CNN acceleration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep CNNs are deployed on general-purpose platforms like GPU, then base calling accuracy is improved, but processing speed and energy efficiency deteriorate due to intensive computation and frequent memory accesses
Solution Approach 1:
The patent replaces the general-purpose GPU computing architecture with a specialized neural network processor architecture that integrates computation and memory closer together. This architectural substitution reduces memory access latency and bandwidth requirements, enabling faster processing while maintaining the deep CNN computational patterns needed for accurate base calling
Solution Approach 2:
The patent changes the computational parameters by implementing custom dataflow patterns and memory access schemes optimized for convolution operations. This includes reorganizing data layouts, using specialized memory hierarchies, and optimizing parallel computation parameters to achieve both high accuracy and fast processing speeds
2Measurement precision
If deep CNNs are deployed on general-purpose platforms like GPU, then base calling accuracy is improved, but energy consumption increases due to intensive computation
Solution Approach 1:
The patent substitutes the energy-intensive GPU architecture with a dedicated neural network processor that uses specialized hardware circuits for convolution operations. This reduces the energy required per computation operation while preserving the deep CNN model's ability to achieve high base calling accuracy
Solution Approach 2:
The patent optimizes energy consumption by changing computational parameters including using lower precision arithmetic where appropriate, optimizing activation functions, and implementing efficient memory access patterns that reduce the energy cost of data movement between memory and processing units
3Productivity
If specialized acceleration hardware like FPGA is used for convolution operations, then processing speed is improved, but device complexity increases
Solution Approach 1:
The patent designs a neural network processor with universal building blocks and modular architecture that can be configured for different neural network models and applications. This multi-functionality reduces the apparent complexity by providing a standardized platform that handles various base calling scenarios without requiring completely different hardware designs
Solution Approach 2:
The patent implements dynamic configuration capabilities that allow the hardware to adapt its structure and parameters based on the specific neural network model being executed. This dynamic reconfigurability optimizes performance for different algorithms while managing complexity through software-controlled hardware adaptation
Data Source
AI summary
A method of base calling using at least two base callers is disclosed. The method includes executing at least a first base caller and a second base caller on sensor data generated for sensing cycles in a series of sensing cycles; generating, by the first base caller, first classification information associated with the sensor data, based on executing the first base caller on the sensor data; and generating, by the second base caller, second classification information associated with the sensor data, based on executing the second base caller on the sensor data. In an example, based on the first classification information and the second classification information, a final classification information is generated, where the final classification information includes one or more base calls for the sensor data.