Repeated isomiRs in the mirtop.tsv output

[bounlu]

The mirtop.tsv output contains repeated rows for some isomiRs with identical first 12 columns:

UID  Read  miRNA  Variant  iso_5p  iso_3p  iso_add3p  iso_snp  iso_5p_nt  iso_3p_nt  iso_add3p_nt  iso_snp_nt
iso-16-03170FE  ACAGTAGTCTACACAT  hsa-miR-199b-3p  iso_3p:-6,iso_snv_central  0  -6  0  1  0  tggtta  0  11AG
iso-16-03170FE  ACAGTAGTCTACACAT  hsa-miR-199a-3p  iso_3p:-6,iso_snv_central  0  -6  0  1  0  tggtta  0  11AG
iso-16-03170FE  ACAGTAGTCTACACAT  hsa-miR-199a-3p  iso_3p:-6,iso_snv_central  0  -6  0  1  0  tggtta  0  11AG

Their counts in the samples differ slightly though. I would expect these isomiRs to be merged and counted together in the output. I understand from mirtop.gff that they come from different precursors with identical mature sequences, but then I don't understand why their counts are different in the same sample?

Is this a bug? If not, what is the reason behind repeated isomiRs with identical sequences but different counts in the same sample?