Indels are common in the human genome and contribute to genetic diversity and human disease (47–49). In the clinical molecular oncology laboratory, the detection of small (defined here as <10 bp) and medium (defined here as >10 but <1 kb) indels is important to many cancers. Of particular clinical significance are the NPM1 insertion, FLT3 internal tandem duplication (FLT3-ITD), KIT exon 8 indels in acute myeloid leukemia, and EGFR exons 19 and 21 insertions and deletions in lung cancer (50–53). While small-and mediumsized indels are usually simple to detect by Sanger sequencing or gel capillary–based sizing methods, indel detection by NGS methods has been challenging largely because of the short read lengths generated by NGS methods. In general, small indels can be called with reasonable sensitivity from NGS data, although the specificity tends to be low. Medium-sized indels, such as the FLT3-ITD, however, have proven difficult to detect by most, but not all, methods (22). Further, most indel detection software is biased to detect deletions over insertions, as inserted sequences are more difficult to align to the reference sequence as described below. Indel detection software can be divided into four major categories, although there is considerable overlap among software packages.
