Shouji: A Fast and Efficient Pre-Alignment Filter for Sequence Alignment (1809.07858v4)

Abstract: Motivation: The ability to generate massive amounts of sequencing data continues to overwhelm the processing capability of existing algorithms and compute infrastructures. In this work, we explore the use of hardware/software co-design and hardware acceleration to significantly reduce the execution time of short sequence alignment, a crucial step in analyzing sequenced genomes. We introduce Shouji, a highly-parallel and accurate pre-alignment filter that remarkably reduces the need for computationally-costly dynamic programming algorithms. The first key idea of our proposed pre-alignment filter is to provide high filtering accuracy by correctly detecting all common subsequences shared between two given sequences. The second key idea is to design a hardware accelerator that adopts modern FPGA (Field-Programmable Gate Array) architectures to further boost the performance of our algorithm. Results: Shouji significantly improves the accuracy of pre-alignment filtering by up to two orders of magnitude compared to the state-of-the-art pre-alignment filters, GateKeeper and SHD. Our FPGA-based accelerator is up to three orders of magnitude faster than the equivalent CPU implementation of Shouji. Using a single FPGA chip, we benchmark the benefits of integrating Shouji with five state-of-the-art sequence aligners, designed for different computing platforms. The addition of Shouji as a pre-alignment step reduces the execution time of the five state-of-the-art sequence aligners by up to 18.8x. Shouji can be adapted for any bioinformatics pipeline that performs sequence alignment for verification. Unlike most existing methods that aim to accelerate sequence alignment, Shouji does not sacrifice any of the aligner capabilities, as it does not modify or replace the alignment step. Availability: https://github.com/CMU-SAFARI/Shouji

• The paper introduces Shouji, a pre-alignment filter that boosts accuracy up to two orders of magnitude over leading methods.
• It utilizes hardware/software co-design with FPGA acceleration to achieve up to three orders of magnitude faster processing than CPU implementations.
• The method reduces computational overhead in sequence alignment by using a scalable sliding window approach to detect common subsequences.

Shouji: Enhancing Computational Efficiency in Short Sequence Alignment

The paper "Shouji: A Fast and Efficient Pre-Alignment Filter for Sequence Alignment" presents a novel approach to address the ongoing challenge of processing the vast amount of sequencing data being generated in genomics. The authors introduce a pre-alignment filter named Shouji, which aims to significantly reduce the computational time associated with sequence alignment without compromising the accuracy and integrity of the alignment process. By leveraging hardware/software co-design and modern FPGA architectures, Shouji stands apart through its innovative filtering mechanism and hardware acceleration, achieving remarkable improvements in both speed and filtering accuracy compared to existing methodologies.

Shouji's approach is founded on two primary concepts: providing high filtering accuracy by detecting common subsequences between sequences, and a hardware accelerator that exploits the parallelism afforded by FPGA platforms to expedite the filtering process. The paper reports that Shouji increases the accuracy of pre-alignment filtering up to two orders of magnitude over GateKeeper and SHD, two of the current leading filters. It also demonstrates that the FPGA implementation of Shouji executes up to three orders of magnitude faster than a CPU-based implementation.

The paper details the methodology of Shouji, beginning with the construction of a neighborhood map—a binary matrix indicating matches and mismatches between sequence pairs. The challenge in identifying non-overlapping common subsequences, which form the core of Shouji's efficiency, is tackled by employing a sliding search window approach that drastically limits computational overhead while maintaining accuracy. This strategy allows for a scalable and parallelizable search process, aligning well with FPGA architectures.

In examining Shouji's impact on sequence alignment, the paper benchmarks the integration of Shouji with five state-of-the-art sequence aligners across different computing platforms. Notably, Shouji reduces the execution time of aligners by as much as 18.8x, underscoring its practical capability to streamline computational tasks in genomics. Additionally, Shouji shows promising results when combined with read mappers like mrFAST and BWA-MEM, though potential exists for further optimization.

The authors also investigate the resource utilization on FPGAs, showing that Shouji requires significantly less hardware resource compared to other systems such as MAGNET, facilitating its implementation across a broader range of FPGA architectures. The paper concludes with an acknowledgment of the importance of specialized hardware in genomic processing and discusses future avenues of development, including privacy-preserving implementations and adaptations for longer sequence alignments.

Overall, Shouji represents a substantial advancement towards efficient genomic data processing, offering both speed and accuracy without sacrificing the quality of sequence alignments. This paper exemplifies how advanced computing techniques can interweave with bioinformatics to tackle the challenges posed by the sheer scale of contemporary sequencing endeavors. Looking ahead, the integration of FPGA-based pre-alignment filtering can be anticipated to play a pivotal role in the evolution of genome analysis workflows, potentially catalyzing real-time and precise genomic analyses.