Figure 2.

Sample cases for different interval offsets. k-mers shared between intervals A and B are underlined. (a) In an ideal partition, the shared k-mers are perfectly captured in both intervals. (b) In a worse-case partition, only half of the shared k-mers are captured in the cell, leading to a misleading identity estimate for this region. (c) By keeping A fixed, but expanding B to $B^{\prime }$, ModDotPlot is able to better capture the similarity between two similar sequences with different offsets. The containment index of A in $B^{\prime }$ is then used to determine the score of the dot plot matrix cell $M_w(A,B)$