A 28nm Fully Integrated End-to-End Genome Analysis Accelerator for Next-Generation Sequencing

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Next-generation sequencing (NGS) generates vast amounts of data requiring efficient analysis pipelines.
• Current bioinformatics tools face challenges in processing speed, precision, and energy consumption for complex genomic analyses.
• End-to-end analysis of short-read data, including mapping, variant calling, and genotyping, is computationally intensive.

Purpose of the Study:

• To develop and present the first end-to-end accelerator for next-generation sequencing (NGS) data analysis.
• To enhance the efficiency and accuracy of short-read mapping, haplotype calling, variant calling, and genotyping.
• To provide a hardware acceleration solution that outperforms existing software and hardware approaches.

Main Methods:

• Utilized FM-index for exact-match short-read mapping and dynamic programming for inexact matching.
• Implemented a rapid similarity calculation and a rescue technique to optimize mapping workload and sensitivity.
• Developed parallel k-mer processing for de Bruijn graph construction and haplotype assembly.
• Integrated variant discovery and genotype likelihood computing engines for comprehensive variant analysis.

Main Results:

• Achieved end-to-end data analysis for the 50× PrecisionFDA dataset in an average of 28.2 minutes.
• Demonstrated a 3-to-59× higher throughput compared to existing solutions.
• Attained high precision (99.79%) and sensitivity (99.03%) in variant calling and genotyping.
• Reported a 935× higher energy efficiency than the Illumina DRAGEN FPGA acceleration system.

Conclusions:

• The developed accelerator provides a significant advancement in NGS data analysis efficiency and performance.
• This hardware solution offers a substantial improvement in throughput, precision, and energy efficiency for critical genomic tasks.
• The end-to-end approach streamlines the complex bioinformatics workflow, enabling faster and more accurate genomic insights.