can some argument change the criteria to obtain high-confidence SV genotype call with low coverage NGS data?

[liuqi-C]

I have a NGS data with coverage 10-15X, and I use the (Manta, CNVnator, Lumpy) + Svimmer + Graphtyper2 to detect SV and genotype SV. I got result which many SV site have LowPratio.

when I see the SI of original paper (Hannes P. Eggertsson et al. 2019), the criteria include:

1. all genotyping models must have at least 10 unique reads

the mean coverage in my data is only 10-15X, maybe many site can't meet this criteria. So, I want to know if there are some arguments I can change the criteria?

Thanks!