Figure 5

Match extension by window alignment for W = 6 and K = 2. Match extension by window alignment for window size W = 6 and maximum number of errors K = 2. A: We show the original pair of sub-sequences. Beginning from the left, the first window with 3 errors is printed in bold. Both neighbouring windows are separated by blanks. The work arrays (1) are initialized with the contents of the left hand side windows. B: We show the changes of the work arrays for the single steps of iterations described in Section Algorithm. Step 1: The central and right hand side blocks of (1) are overwritten by the next 12 original sequence characters yielding (2). An optimal alignment of these 12 sequence characters is given in (3). Step 2: The central blocks of (2) are overwritten by the first six characters of the alignment (3) and the right hand side blocks of (2) are overwritten by the next 6 original sequence characters yielding (4). Step 3: From the left to the right the first window of size 6 with 3 errors is printed in bold. Its starting position is T = 2W +1 and the existence of gaps in the central block requires step 4. Step 4: The central blocks of (5) are shifted to the start positions of the work arrays to set the initialization for the next round of iteration shown to the left of the first round. The third round of iteration ends in step 3 without gaps in the central and right hand side blocks of (5) and initiates match extension without gaps as described in the text.