Acceleration Of Nucleotide Semi-Global Alignment With Adaptive Banded Dynamic Programming [article]

Hajime Suzuki, Masahiro Kasahara
2017 bioRxiv   pre-print
Pairwise alignment of nucleotide sequences has previously been carried out using the seed-and-extend strategy, where we enumerate seeds (shared patterns) between sequences and then extend the seeds by Smith-Waterman-like semi-global dynamic programming to obtain full pairwise alignments. With the advent of massively parallel short read sequencers, algorithms and data structures for efficiently finding seeds have been extensively explored. However, recent advances in single-molecule sequencing
more » ... chnologies have enabled us to obtain millions of reads, each of which is orders of magnitude longer than those output by the short-read sequencers, demanding a faster algorithm for the extension step that accounts for most of the computation time required for pairwise local alignment. Our goal is to design a faster extension algorithm suitable for single-molecule sequencers with high sequencing error rates (e.g., 10-15%) and with more frequent insertions and deletions than substitutions. We propose an adaptive banded dynamic programming algorithm for calculating pairwise semi-global alignment of nucleotide sequences that allows a relatively high insertion or deletion rate while keeping band width relatively low (e.g., 32 or 64 cells) regardless of sequence lengths. Our new algorithm eliminated mutual dependences between elements in a vector, allowing an efficient Single-Instruction-Multiple-Data parallelization. We experimentally demonstrate that our algorithm runs approximately 5x faster than the extension alignment algorithm in NCBI BLAST+ while retaining similar sensitivity (recall). We also show that our extension algorithm is more sensitive than the extension alignment routine in DALIGNER, while the computation time is comparable.
doi:10.1101/130633 fatcat:ostn3u6oejclbanwznilspczde