Fig. 3

An example of the two-pass query optimization on a set of 8 query reads (1), that are assigned ids from 1 to 8. First, the color-id lists are generated from the input reads (2), then sorted and collated (3) so that the set of read ids having the same color-id list is retained for each distinct list. For example, the reads 1, 4, and 7 all have the same color-id list [2, 4, 13, 25]. Lastly, intersections are performed for each color-id list and each original read is annotated with its result (4). We have four distinct color-id lists in the example, hence four results \(R_1\), \(R_2\), \(R_3\), and \(R_4\) are computed. In the picture, the black thin arrows from (3) to (4) logically show how each original read is associated to its result \(R_i\), \(i=1,\ldots ,4\)