High-performance hardware implementation of a parallel database search engine for real-time peptide mass...

Area of Science:

• Biochemistry
• Computational Biology
• Bioinformatics

Background:

• Peptide mass fingerprinting (PMF) is a crucial technique for protein identification.
• Current PMF methods involve mass spectra processing and database searching, which can be time-consuming.
• Real-time protein identification could revolutionize PMF applications.

Purpose of the Study:

• To develop a hardware-based solution for real-time protein identification using PMF.
• To significantly accelerate the database search stage of the PMF process.
• To present a complete real-time PMF solution by combining hardware processors.

Main Methods:

• Designed and implemented a parallel database search engine using Field Programmable Gate Array (FPGA) hardware.
• Utilized a Xilinx Virtex 2 FPGA operating at 100 MHz for the database search engine.
• Integrated the database search engine with a previously developed hardware mass spectra processor.

Main Results:

• Achieved an 1800-fold speed-up for the database search engine compared to a conventional C software routine.
• Demonstrated the scalability of the FPGA design for further performance enhancement.
• The combined hardware processors offer a complete real-time PMF protein identification solution.

Conclusions:

• Hardware acceleration, particularly using FPGAs, dramatically improves PMF processing speed.
• The developed parallel database search engine is a key component for real-time protein identification.
• This integrated hardware solution has the potential to transform the deployment and use of PMF.