What is Indelocator?
Indelocator is a software tool for calling short indels in next generation sequencing data.
How does Indelocator work?
Indelocator is mainly the engine for collecting various read count and alignment quality-related statistics around putative indel events. It does not perform realignment or split-read alignment, but relies on the alignments already recorded in the input bam file(s). It is thus very important to use well-aligned input bam files, as misaligned reads will not be corrected and will not be counted towards their correct allele. Running local realignment around indels algorithm from GATK is highly recommended prior to using Indelocator.
In its default mode, Indelocator uses two input bam files (or two sets of input files), for normal and tumor samples. Events present in tumor sample above the specified thresholds and having sufficient coverage in normal sample (another input parameter) are recorded in the output alongside with various alignment metrics, and are also annotated as (putatively) germline or somatic depending on the presence of the alternative allele in normal sample (see GATK documentation for more detail).
Indelocator also supports single-sample mode, in which it tallies same alignment metrics around each putative indel event in a (single) sample.
Output of Indelocator is a high-sensitivity list of putative indel events containing large numbers of false positives. The statistics reported for each event have to be used to custom-filter the list in order to lower false positive rate.