filter option should not be taken into account when computing processed_alignment key in store
If we run a first scan with a filter option, and then run a second one, we need the processed alignments in run 1 to be retrieved when running the second time. But if the hash key is computed using the filter option (as it is now the case) the hash key is different.