DNA Sequence Alignment

Spring 2013

John Greth
Onur Mutlu
Project description

DNA Sequence alignment is the process in which small DNA sequences are mapped to their correct location in the reference genome. This process is very computationally expensive: it takes an average desktop computer weeks to months to sort through a human's entire DNA. It is also very expensive (a single DNA test can cost a patient $1000). As medical researchers find more and more uses for dna tests, it is becoming more important that we find a way to make the sequencing process faster and less expensive.

The first step to DNA sequence alignment is filtering: narrowing the mapping scope down to only a few possible locations in the reference genome. My objective is to speed up DNA sequencing by implementing the first step on an FPGA (Field Programmable Gate Array). FPGAs are more expensive than computer processors, but they can be faster for some computations because they are programmable at the logic/gate level. In order to meet the objective, I am modifying existing sequencing algorithms to optimize for the FPGA platform, running simulations to evaluate expected performance increase, implementing the modified algorithm on an FPGA, and benchmarking the algorithm to verify that it has a high speed to cost ratio.

Return to project list