A Storage-Efficient Suffix Tree Construction Algorithm for Human Genome Sequences

Abstract

The suffix tree is one of most widely adopted indexes in the application of genome sequence alignment. Although it supports very fast alignment, it has a couple of shortcomings, such as a very long construction time and a very large volume size. Loh et al. [7] proposed a suffix tree construction algorithm with dramatically improved performance; however, the size still remains as a challenging problem. We propose an algorithm by extending the one by Loh et al. to reduce the suffix tree size. As a result of our experiments, our algorithm constructed a suffix tree of approximately 60% of the size within almost the same time period.