Hylitech, Inc. proposes a solution to overcome a major computational bottleneck in proteomics-by building a dedicated low-cost supercomputer system that performs real time peptide and protein identification from mass spectral data. Current peptide and protein identification requires computational processing that lags mass spectral data acquisition by hours or even days. The utilization of an FPGA platform will allow matches between experimental data and possible peptides or modified peptides to be evaluated in real time, while data acquisition is in progress. Fast identification of peptides and modified peptides will allow more efficient use of scarce mass spectrometer time and enable a much deeper investigation to discover low abundance regulatory species in the proteome. This time and cost effective approach to proteomic data analysis will speed research, improve discovery of disease markers, facilitate development of more effective diagnostics, and foster the creation of personalized medicine. Real time analysis will be accomplished by designing a peptide processing unit (PPU) that will be replicated across FPGAs. A system level design will achieve the supercomputing performance required to provide a real-time solution. This will bring major acceleration of proteomic data analysis to biomedicine, while providing significant cost savings compared to conventional multi-microprocessor based systems.
| Snider, Ross K (2007) Efficient calculation of exact mass isotopic distributions. J Am Soc Mass Spectrom 18:1511-5 |
