Enhanced Suffix Trees for Very Large DNA Sequences
Recent advances in bio-technology have provided rapid accumulation of biological DNA sequence data. New techniques are required for fast, scalable, and versatile processing of such data. Suffix tree (ST) is a data structure used for indexing genome data. This, however, comes with a price: it occupi...
| Main Author: | |
|---|---|
| Format: | Others |
| Published: |
2011
|
| Online Access: | http://spectrum.library.concordia.ca/35799/1/Fan_MCompSc_F2011.pdf Fan, Si Ai <http://spectrum.library.concordia.ca/view/creators/Fan=3ASi_Ai=3A=3A.html> (2011) Enhanced Suffix Trees for Very Large DNA Sequences. Masters thesis, Concordia University. |
| Summary: | Recent advances in bio-technology have provided rapid accumulation of biological DNA sequence data. New techniques are required for fast, scalable, and versatile processing of such data.
Suffix tree (ST) is a data structure used for indexing genome data. This, however, comes with a price: it occupies a space that is about 10 times more than the input size. Existing disk-based ST index techniques either suffer from data skew problem, like TDD and HST, or are not space efficient for very large sequences, like TRELLIS and B2ST. We propose a new disk-based ST index, called Compact Binary Suffix Tree (CBST), together with a construction algorithm, which can support DNA sequences of size up to 256 terabyte. The results of our numerous experiments indicated that, compared to existing ST and suffix array techniques, CBST is superior in speed, space requirement, and scalability. It is the fastest among the disk-based techniques for very large sequences.
|
|---|