Overlapping VCF entries in repetitive regions
I'm trying to use GATK to genotype short tandem repeats, but am a bit confused as to how I should interpret the VCF output. When I run HaplotypeCaller, I quite frequently obtain VCF entries in which one variants's reference allele extends into the next variant. For instance, the first five columns in such an instance may look like the following:
1 32826477 . AGATAGATAGATAGATAGAT A
1 32826481 . AGATAGATAGATAGAT A
In this case, the first record extends to position 32826496 and so overlaps with the latter record. If I had a sample with phased genotypes 0|1 at the first record and 0|1 at the second record, obviously this would generate some problems as the same set of bases are being deleted twice. How does the HaplotypeCaller ensure that these records are self-consistent and how should I interpret these overlapping records? Is there a way that I can force the HaplotypeCaller to combine overlapping records and report them as a single variant?
Thanks so much for the help!