A Suffix Tree Construction Algorithm for DNA Sequences

The suffix tree is a powerful data structure in string processing and DNA sequence comparisons. However, constructing suffix trees being very greedy in space is a fatal drawback. In addition, the performance of the suffix tree construction using suffix link will rapidly degrade with the increase of the scale of sequences to be handled because of the random access. In order to overcome these disadvantages, a new bit layout is used for the nodes of a suffix tree which has less space requirements. Based on this an algorithm to construct suffix tree for DNA sequences is proposed using partitioning strategies. The effectiveness for the proposed algorithm is shown in the testing cases from NCBI web site. Comparisons with Kurtz's algorithm in space requirements and running time have been made in the experiments. The results show that the proposed algorithm is memory-efficient and has a better performance over Kurtz's algorithm on the average running time.