miRDeep2 sample read count different between mapper.pl and quantifier.pl

Question

[sg197]

Hi,

I'm using miRDeep2 for novel microRNA prediction and also counting reads for both novel and known microRNAs. After using mapper.pl, I get this output:

#desc   total   mapped  unmapped    %mapped %unmapped
S01: 24882381   22106271    2776110 88.843  11.157
S02: 30468230   26078475    4389755 85.592  14.408
S03: 16914527   14834666    2079861 87.704  12.296
S04: 22149373   19447606    2701767 87.802  12.198
S05: 21958296   18236811    3721485 83.052  16.948
S06: 22475098   19439477    3035621 86.493  13.507

After I use miRDeep2.pl to do novel prediction, I then use qauntifier.pl to map reads to both known and predicted novel microRNAs (from miRDeep2.pl), this is the output:

#desc   total   mapped  unmapped    %mapped %unmapped
S01: 24805222   20922665    3882557 84.348  15.652
S02: 30369347   24465997    5903350 80.561  19.439
S03: 16859122   14137153    2721969 83.855  16.145
S04: 22076946   18444621    3632325 83.547  16.453
S05: 21886410   16937671    4948739 77.389  22.611
S06: 22401867   18424644    3977223 82.246  17.754

I'm confused why for each sample, the total number of reads is different in each output. The number of total reads in the top panel matches the number of reads from the fastqc reports, so I'm worried I'm losing reads somewhere along the way (maybe when collapsing fasta reads in mapper.pl) and thats why theres fewer in the second panel? Any help appreciated! Thanks