This system reduces the time required to identify a protein from mass spectrometry data from tens of minutes to less than 100 milliseconds. This is achieved by running bioinformatics algorithms on highly optimised digital processors that are programmed to run on a 'reconfigurable chip', the FPGA (Field Programmable Gate Array).

An FPGA is a type of digital integrated circuit that can be reprogrammed after it is manufactured. The FPGA devices sit on a reconfigurable card that is interfaced with a PC via the standard PCI bus. The card can be configured to have 4 FPGA devices and 3Gbytes of RAM memory to store entire protein databases. The processing speed of this system allows the user to get a match to a database entry in real-time.

The platform can be easily adapted to deal with data generated by other instruments (LS-MS/MS), for which data processing can take even longer. The advantages of such a system can be summarised as follows:

• Delivers the performance of hundreds of CPUs in a single PC box
• Can be installed on servers running Linux or Windows requiring no additional power or cooling
• No “per CPU” licensing costs associated with a cluster system
• Much higher up-time compared with clustered solutions
• Direct fast Ethernet connection with Mass Spectrometers
• On-Board proprietary protein database implementation enables fast & parallel searches

The digital hardware designs that run on the FPGA system are protected by copyright. The implementation is distributed in the form of a bitstream (the equivalent of the executable file on a computer). The bitstream contains all the information required to configure the FPGA with the specific design. The manufacturer of the devices used in our system supports advanced Triple DES (Data Encryption Standard) security embedded in them. This means that the user only has access to encrypted bitstreams that will be decoded (given the correct key) by dedicated on-chip logic making it virtually impossible to duplicate or reverse-engineer the FPGA solution. The encryption facility can also be used to implement different pricing models- different keys enable access to different features of the design.

The raw spectra processor which generates the Protein Mass Fingerprint has been validated using an artificial QconCAT protein chosen designed so that all tryptic fragments fell within the range 1000–2500 m/z. Deisotoping performance was also tested separately through a deamidation experiment involving glyceraldehyde 3 phosphate dehydrogenase. Data was processed in parallel with the MassLynx software for comparison. The database search engine was tested using both synthetic and experimental data.

• Real-time protein identification

• Ideal solution for performing batch analysis

• Bioinformatics, proteomics, drug discovery, systems biology, synthetic biology, biomarker identification applications

• Add-on system suitable for interface with current MS instrumentation