Designing seeds for similarity search in genomic DNA

This work addresses the problem of designing a seed to optimize performance of seeded alignment. We give a fast, simple algorithm based on finite automata for evaluating the sensitivity of a seed in a Markov model of ungapped alignments, along with extensions to mixtures and inhomogeneous Markov models. We give intuition and theoretical results on which seeds are good choices. Finally, we describe Mandala, a software tool for seed design, and show that it can be used to improve the sensitivity of alignment in practice.