Succinct data structures were designed to store and/or index data with a relatively small alphabet size, a rather skewed distribution and/or, a considerable amount of repetitiveness.
The work and depth of our first parallel wavelet tree algorithm match those of the best existing parallel algorithm while requiring asymptotically less memory and our second algorithm achieves the same asymptotic bounds for small alphabet sizes.
The simplified version is more efficient for DNA sequences with a very small alphabet.
Some of the motifs were also modified and the modified versions of the motifs were labeled with an additional small alphabet.
In recent years, researchers in databases, both multi-media and text, investigated indices in high-dimensional spaces (Böhm et al., 2001; Bustos and Navarro, 2009; Houle and Sakuma, 2005; Navarro and Chávez, 2006; Yao et al., 2010), but the small alphabet size of DNA that leads to non-sparse frequency vectors preclude their use here.
The answer is positive; thanks to the small alphabet the search times (the count query, in which we return the number of matches only) of the FM-index, in the best current implementation, may be comparable with the suffix array ([ 86], Table Six and Seven).
This is not observed, suggesting that in practice smaller alphabets do not necessarily improve statistical performance.
Much previous work, including [ 5] which used the same myoglobin alignment and tree, has focused on the assumption that smaller alphabets are generally better.
Thirdly, we show how to compute only the size of the smallest alphabet in O(m) time.
This is the process of converting an arbitrary random source into a source with smaller alphabet and almost uniform distribution.
