Long spaced seeds for finding similarities between biological sequences

Homology search finds similar segments between two biological sequences, such as DNA or protein sequences. A significant fraction of the computing power in the world is devoted to finding similarities between biological sequences. The introduction of optimal spaced seeds in [Ma et al., Bioinformatics 18 (2002) 440–445] has increased both the sensitivity and the speed of homology search and it has been adopted by many alignment programs such as BLAST. In spite of significant amount of work, there are no algorithms able to compute long good seeds. We present a different approach here by introducing a new measure that has two desired properties: (i) it is highly correlated with sensitivity of spaced seeds and (ii) it is easily computable. Using this measure we give algorithms that compute better seeds than all previous ones. The fact that sensitivity is not required is essential as it enables us to compute very long good seeds, far beyond the size for which sensitivity can be computed.