A Design of a Hybrid System for DNA Sequence Alignment

This paper describes a parallel algorithm and its needed architecture and a complementary sequential algorithm for solving sequence alignment problem on DNA (Deoxyribonucleic acid) molecules. The parallel algorithm is considered much faster than sequential algorithms used to perform sequence alignment; the initialization operation is done by activating a number of processing elements each compares the two sequences simultaneously and weights this comparison between each nucleotide from the first and the second DNA sequences. Then, the sequence matching operation is performed also simultaneously on the same processing elements. Both the initialization operation and the sequence matching operation are done in only two clock cycles. The proposed sequence matching operation is considered a new approach that highlights the subsequences matched between the given two DNA sequences. This operation provides a good indication for linking the matched subsequences of the two DNA sequences depending on a threshold K. which indicates the number of subsequences of highest score that can be linked. A parallel architecture is also presented as the algorithms are performed using it. A simple sequential algorithm is then presented to get the final alignment between the two DNA sequences. This hybrid system is considered a step towards a complete parallel processing architecture to solve computationally intensive applications of DNA.