A variable length hash method for faster short read mapping on FPGA

摘要: Short read mapping is a process to align the short reads, which are fixed-length fragments of target genome, given reference genome identify mutations in genome. Because rapid development Next Generation Sequencing (NGS) technologies, faster required. In this paper, we propose variable length hash method further accelerate FPGA systems. hash-based algorithms, fixed sub-string each read, called seed, used as key. However, many different seeds mapped into same slots because high ununiformity human and fruitless key comparisons performed. To equalize slot size, an optimized function that changes bit masks adaptively. With approach, it possible improve performance all systems based on functions. The for comparison our system Xilinx XC7VX690T XC6VLX240T can be improved two-times, total outperforms any existing