Running Joint Calling

[HenryMarshall97]

I have tumour biopsies and matched blood. From the whole biopsies, I have cut 10-16 smaller samples, and then undertaken WGS on smaller samples and whole blood. I find when I run joint calling using normal.bam followed by the 10-16 .bam files, the number of SVs called is significantly lower than when running each sample individually against matched normal blood. The number of SVs called after running joint calling is incongruence with what would be expected just based off CNV analysis (i.e some have 0 in a genome that demonstrates multiple LOH and chromothripsis). Any suggestions / advice?