Fig. 4

Timings of 1 million nucleotide queries using a partial k-mer table of length 1 (blue), and of length 12 (orange). A suffix array compression ratio of 4 was used for each index. The performance benefits of the partial k-mer table are most obvious in the count() function, whereas the performance benefit in the locate() function is most notable for longer queries that generate fewer hits (i.e., when the number of backtrace steps is relatively small)