TL;DR: This work estimated the confidence intervals/thresholds of the popular algorithms at the genomic region level and increased the detection power of these algorithms by at least 5 folds, which should drastically reduce the sequencing cost of BSA-Seq studies. (via Semantic Scholar)