Fpga-based accelerator for the generation of pseudo-amino acid composition
Loading...
Date
2015-08-01
Authors
Ching Chee Chow
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
One of the biggest challenges in protein prediction post genomic age is narrowing the gap between the number of newly discovered and uncharacterized proteins and the number of known proteins in protein data banks. This leads to increased demand for efficient techniques to accurately predict protein attributes based solely on its sequence-order information. The Pseudo-Amino Acid Composition (PseAAC) is a modeling technique that incorporates, selectively, sequence-order information of a protein into a discrete model. PseAAC has been applied in numerous protein-related researches using various software-based PseAAC generators. Since this often involves large-scale data processing, computation time is of the essence. The prospect of further reducing computation time of the software is limited due to the sequential nature of software execution. Alternative platform such as programmable hardware has emerged as a solution to this bottleneck. Programmable hardware such as Field Programmable Gate Array (FPGA) enables parallel processing that speeds up computation of PseAAC. In this research, an FPGA-based PseAAC generator architecture is proposed. The architecture consists of several modules. To speed up computation, the two most computation-heavy modules of the architecture, the Sum-of-Small-T and T-u-minus-20, are designed to run in parallel. The generator is realized on the Altera Cyclone III FPGA and achieves computation speed increase of up to 31.5 times over a Perl-based PseAAC generator. In conclusion, significant computation speed improvement is achieved by designing the PseAAC generator to capitalize on the parallel processing capability of the FPGA.