非表示:
キーワード:
-
要旨:
Suffix-trees are popular indexing structures for various sequence processing
problems in biological data management. We investigate here the possibility of
enhancing the search efficiency of disk-resident suffix-trees through
customized layouts of tree-nodes to disk-pages. Specifically, we propose a new
layout strategy, called Stellar, that provides significantly improved search
performance on a representative set of real genomic sequences. Further, Stellar
supports both the standard root-to-leaf lookup queries as well as sophisticated
sequence search algorithms that exploit the suffix-links of suffix-trees. Our
results are encouraging with regard to the ultimate objective of seamlessly
integrating sequence processing in database engines.