[Questions] Does including a few trios in the dataset improve phasing? What is the fate of missing data?

[alxsimon]

Hi,
I'm trying to use Shapeit5 on a non-model species and I have a few questions:

1. In addition to the main dataset of ~200 individuals, I have access to 3 trios. I am wondering if including them in the dataset and using the --pedigree option is expected to slightly improve the phasing accuracy?

2. I'm puzzled by the fact that the final output (after phase_common and phase_rare) does not contain missing data, I thought Shapeit was not doing the imputation? Are there some imputation and are markers with missing data dropped?
   In my test I go for instance from 9336548 snps to 8553795, and I'd like to understand why. Are those snps just dropped because phase cannot be inferred accurately?

Thanks in advance for your time.
Alexis