Figure 2

Automated identification of sister chromatid exchange (SCE) from Strand-seq data. (a) Gross directional mapping data are thresholded to remove bins with unexpectedly high or low read numbers, and analyzed using DNAcopy. Inherited template numbers are converted to a value between 1 and -1 for DNAcopy to make only one of three calls: WW, WC, or CC. DNAcopy defines an interval across two bins, so with a bin size set to 200 kb, the SCE event will be located to within 400 kb. (b) Localization is then iterated by subdividing the identified region into bins one-fifth of the original size (80 kb on first iteration), and re-running DNAcopy. A single bin size is used as padding to aid detection of SCE events at bin boundaries. The iterations of re-running DNAcopy continue until less than 50 reads remain within the interval. (c) A second algorithm identifies the first read to map in a different direction (W read at chr13:19,203,283), then performs a check that the 10 preceding reads are all in the expected direction (10 C reads), and at least 20% of succeeding reads are in the other direction. The interval is refined to a distance between two reads. Abbreviations: C, Crick; W, Watson.