I was running the haplotypeCaller for many samples, but some variants (validated as true positives by using other techniques) within these samples are not called by the haplotypeCaller. I saw in the bam files that most of these variants are located on the outside of duplicated reads (around 200 reads). Most of my data consists of duplicated reads. First I thought that the duplicated reads were filtered out by the read filters which are automatically applied (like duplicateReadFilter), but when I checked it this was not the case. I was wondering why my true variants are not called by the HaplotypeCaller and if there is an option to resolve this problem?