MGI gene sequencing company

Change Region
Location:Home / Products
Product Description
MegaBOLT bioinformatics analysis accelerator (hereinafter called MegaBOLT) is an MGI self-developed and MPS-concentrated hardware accelerating system for bioinformatics analysis. MegaBOLT supports the analysis of Whole Genome Sequencing (WGS), Whole Exome Sequencing (WES), and Panel Sequencing on Germline or Somatic data. It is 20 times faster than the traditional approach. Fast, integrated, easy to use, and economic are MegaBOLT’s most prominent features.
MegaBOLT products include the workstation and rack server, which can be applied to various scenarios of all scales.
  • Laboratories
  • Hospitals
  • Research institutes
  • Enterprises
Fast
MegaBOLT adopts CPU and FPGA heterogeneous computing to accelerate analysis. It accelerates algorithms such as SOAPnuke, Minimap2, BWA, GATK HaplotypeCaller + MuTect2 through multi-stream and highly parallelized computing architecture, which contributes to a 10X to 20X speedup in massive data analysis.
01.Speed comparison
Germline
Tradition
MegaBOLT
  • WGS 30X NA12878 DNBSEQ
  • WES 400X NA12878 DNBSEQ
Somatic
Tradition
MegaBOLT
  • WGS 80X NA12878_Sim DNBSEQ
  • WES 800X NA12878_Sim DNBSEQ

* The BASIC pipeline only includes the analysis steps from clean data to variant calls, while the FULL pipeline includes additional steps of quality control, statistics, and reports.


Germline: The duration of 30X WGS analysis is shortened from 40 hours to about 2 hours, and that of 400X WES analysis is shortened to about 30 minutes.
Somatic: The duration of 80X (40X Tumor + 40X Normal) analysis is shortened to about 5 hours, and that of 800X (400X Tumor + 400X Normal) WES analysis to about 40 minutes.
02.Multi-task scheduling system acceleration
Integrated with the self-developed multi-task scheduling system, MegaBOLT supports simultaneous multi-task computing on one single server. The computing efficiency is further improved by 16% to 36%.
Germline
MegaBOLT
MegaBOLT-scheduler
  • WGS 30X NA12878 DNBSEQ
  • WES 400X NA12878 DNBSEQ
* The mean value is calculated by executing the same 4 tasks at the same time.
Precision
01.Precision
After speedup, the analysis precision remains steadfast. MegaBOLT is also configured with a deep learning variant calling module, MegaBOLT-DV, which is optimized by algorithms and specified neural network model training. When MegaBOLT is used in combination with PCR-Free library preparation and MGI DNBSEQ™ sequencing technologies, the human WGS variant calling precision can reach SNP 99.9% and INDEL 99%.

02.Deep learning making sequencing more cost-effective
The deep learning variant calling module, MegaBOLT-DV, helps to greatly reduce the sequencing depth and cost without affecting the analysis precision.
Compared with the traditional variant calling results of 30X PCR data, SNP of 15X PCR-Free data from MegaBOLT-DV is close while INDEL better.


Tradition
MegaBOLT-DV
Higher Performance MegaBOLT-Pro
MegaBOLT uses multi-task scheduling system and higher performance machine for large throughput sequencers such as DNBSEQ-T7. Advanced version MegaBOLT-Pro was launched. The performance was further improved by 5 times, achieving the entire process analysis in an average of 30 minutes. Up to 90 times faster than the traditional algorithm.
Germline
Tradition
MegaBOLT
MegaBOLT-Pro
  • WGS 30X NA12878 DNBSEQ
* The FULL pipeline includes additional steps of quality control, statistics, and reports.
* The mean value is calculated by executing the same 8 tasks at the same time.
Integrated
Integrated with the self-developed laboratory information management system ZLIMS, MegaBOLT realizes fully automated sequencing and analysis. It can work together with MGI DNBSEQ™ sequencers or sequencers of other platforms.
  • Laboratory device management and real-time monitoring
  • Sample information recording and tracking
  • Data transmission and management
  • Analysis task scheduling and computing
  • Analysis result/report management and visualization
  • Easy to use
    The Web design makes MegaBOLT simple and easy to use. You can just click a couple of buttons to submit a task and check the results on the website. In this case, a bioinformatic technical background is no longer necessary.
  • Economic
    With a prominent speedup in computing and rigid control over hardware prime cost, a MegaBOLT server works like 10 normal computers in total and is without doubt highly cost-effective.
Heterogeneous computing system of CPU and FPGA
accommodates high-speed sequence analysis
MegaBOLT’s heterogeneous computing system of CPU and FPGA, fulfilling from FASTQ input of sequencing data to BAM output of alignment result and VCF/gVCF output of variant calling result. The FPGA hardware logic circuit is specially designed for bioinformatics computing to increase its analysis speed. We adopt a multi-stream computing architecture for FPGA and divide data into smaller granules for higher efficiency. We also further streamlined the whole process and developed a highly parallelized computing architecture, thus realizing highspeed WGS/WES analysis and computing.

Compared with the common analysis workflow, MegaBOLT stands out for the following features:
  • Specialized FPGA hardware logic circuit for bioinformatics computing
  • Multi-stream computing architecture
  • Highly parallelized computing architecture
  • Cloud computing
    MegaBOLT integrates the cloud computing technology and is launched on the BGI Online cloud platform, which provides more flexible and efficient solutions for bioinformatics analysis. For details, please visit: https://www.bgionline.cn/fpga.html

  • Diverse applications
    MegaBOLT is a comprehensive platform that embraces all sorts of downstream software and third-party applications.
    Available functions
    • Structural variant discovery
    • Copy number variant discovery
    • Variant annotations
    • Simplified genome RAD-Seq variant discovery
    Available applications
    • Inherited disease genetic data analysis and reporting
    • Tumor analysis and targeted drug therapy
    • Metagenomics analysis and interpretation
    • Supporting data for clinically actionable insights
    • Reproductive health analysis and interpretation
Related Documents