Fig. 4

A shows a problem caused by the repetitiveness of the reference genome. In the sequenced genome, one of the occurrences of \(B\) is missing. The two seeds \({m}_{1}\) and \({m}_{2}\) that overlap on the y-axis (top) corresponds to the erroneous entry \(a\) (bottom). Our overlap-elimination technique (as described in the methods section) resolves this problem by shortening \({m}_{1}\) and \({m}_{2}\) to the line \({p}_{cut}\). The result of this shortening is the correct entry \(a{\prime}\). B Two 1 nt deletions between the reference genome and sequenced genome create the small seed \(m\) (top-left). \(m\) is assumed to be deleted by the occurrence filtering of our approach (top-right). This deletion leads to the replacement of the two correct entries \(a\) and \(b\) (bottom-left) with a wrong entry \(c\) (bottom-right). For salvaging such lost seeds, we rediscover \(m\) via a reseeding within the gray boxed area (top-right). C visualizes our reseeding technique for a larger area, where each box corresponds to a recursive call and the gray’s darkness indicates the recursion depth