I have 2 patients, each of which has normal, primary, and metastatic samples whole exome sequenced.
If I compare somatic mutations in primary and metastatic samples
Patient 1:
N_somatic_mutations_primary N_somatic_mutations_metastatic N_intersection
532 778 132
Patient 2:
primary sample has even more somatic mutations than metastatic sample.
Ideally, all somatic mutations in primary should be included in metastatic from same patient. There can be errors, and we could be missing some mutations.
So, there are a few things that can be happening, but you do expect some mutations to be present in the primary and not the met that were real and true mutations. Remember that the tumour is heterogeneous, and depending on what type of tumour we are talking about, some are more heterogeneous than others. When you are sequencing bulk tumour you are looking at that whole mix of mutations as a population. So the clone that founded your metastatic tumour site takes just a fraction of that diversity. It is akin to a population bottleneck in population genetics.
So a few things you want to look at will be your acceptance/filtration criteria for your mutations and perhaps some issues to do with the amount and quality of DNA and the quality of your sequencing. A lot of the mutations that vanished between the primary and met may be low quality for instance, and probably should have been filtered out initially.
