GenotypeGVCFs - Best genotyping strategy for trios

Hi the team,

I am studying de novo variants in trios (child, father, mother). I used a classical pipeline HaplotypeCaller (GVCF mode)/GenotypeGVCFs, and used about 30 BAMs to perform the genotyping.

Then, I used a script to extract the de novo variants in each child, and manually curated them in IGV to eliminate artefacts - generally, variants also present in one or both parents, but at lower frequency that in the child.

My question concerns a reflexion about the best genotyping strategy. Using all the batch, is there a risk to dilute low frequency variants ? Isn't it more pertinent to genotype each trio independently in order to test transmission hypotheses ?

Thanks by advance for your advice,
BPR