Searching by index for similar sequences: the SEQR algorithm

This paper describes a method to efficiently retrieve protein database sequences similar to a query sequence, while allowing for significant numbers of mutations. We call this method SEQR for SEQuence Retrieval. This approach increases the speed of sequence similarity searches by an order of magnitude compared to conventional algorithms at the expense of sensitivity. Furthermore, retrieval time increases less than linearly with the number of sequences, a desirable property during an era when next generation sequencing technologies have yielded greater than exponential increases in sequence records. The lower sensitivity of the algorithm for distantly related sequences compared to benchmarks is not intrinsic to the method itself, but rather due to the procedure used to construct the indexing terms, and may be improved. The indexing terms themselves can be added to standard information retrieval engines, enabling complex queries that include sequence similarity and other descriptors such as taxonomy and text descriptions.