We applied a sliding indo test to identify anomalous regions that contain read pairs significantly different from the entire genome. By default, BreakDancerMini using a fixed indo size of w = μ + 3σ - 2l bp and a step size of 1 bp, here μ and σ are the mean and the standard deviation estimated from the separation distance of normally and confidently (mapping quality > 40) mapped read pairs, and l is the average read length. A to-sample Kolmogorov–Smirnov (KS) test statistic26 is computed for each indo, here Fn(x) and Fn′(x) are the empirical cumulative distribution function (ECDF) estimated from the normal reads in the indo and in the entire genome respectively, and n and n′ are the number of reads in each set; x is the separation distance from 1 bp to a maximum size (∼300 bp); sup denotes the supremum of the set. Obviously, Dnn′ objectively measures the difference between the to ECDFs in terms of both location and shape. To model alignment orientation, we computed to statistics D+nn′ and D-nn′ per indo using reads that are mapped to the plus and the minus strands respectively. A genomic region is classified as anomalous in either the plus or the minus orientation if the corresponding KS statistic exceeds a user-selected threshold. Overlapping anomalous regions in the same orientation are filtered and only the highest scoring one is kept. For small indels, the anomalous regions that support the same variant are required to be in the opposite orientations. In principle, this approach works with any insert size distribution and does not require any predetermined cutoff on the separation distance.