Fig. 1

INDELseek algorithm as illustrated by the BRCA2 complex indel of sample 2. Left: INDELseek directly reads NGS read alignments in the standard SAM/BAM format. After refining matches and mismatches in the supplied alignments, clusters of closely spaced mismatches, insertions and/or deletions are identified as potential complex indel calls. False positives are removed according to filters based on read base quality, allele frequency and allele sequencing depth. Final complex indel calls are reported in the standard VCF format. Right: A representative BWA-MEM alignment of a sample 2 NGS read was shown. The corresponding reference sequence (chr13:g.32912956_32912969) and base calls of the read were shown above the below the alignment, respectively. In the alignment refinement step, M operators were refined as matches (=) and mismatches (X). A cluster of closely spaced variants was identified as a potential complex indel call (highlighted as a red box). The complex indel call passed the defined quality thresholds and was reported as a variant call in VCF format, which corresponds to the BRCA2 complex indel c.4467_4474delinsTGTTTTT