Differences in adapter clipping between CLC and FastXtools

[PPTort]

We are analyzing a miRNA samples and we performed the analysis with CLC genomics and FastXtools.
One intriguing observation is that with the same dataset when perfoming the adapter clipping, CLC removes more than 29 million reads from 48 million, while FastXtools removes only 3 million.

After adapter, quality and size clipping we end up with 11 and 33 million reads!
Sounds too much of a difference for me. Any hint why?
Thanks in advance....

PEPE

PS. Parameters are as follows...

For CLC
Ambiguous nucleotides : max 2 allowed
Adapter seq. (AGATCGGAAGAGCACACGTCT) strand +, action remove, score (2:3:13:4)
minumun length 15
IMPUT READS TRIMMED DISCARDED
trim on quality 48.385.045 1791754 35213
ambiguitiy trim 48349.832 4886 0
adapter trim 48.349.832 11.084.232 29.128.544
length filter 19.221.288 0 8.110.018
OUTPUT 11.111.270

FOR FASTX TOOLS

CRUDE READS 48.385.045
CLIPPING ADAPTER
Adapter seq AGATCGGAAGAGCACACGTCT
too short reads 11.254.859
adapter only 3.093.352
N containing 53.729
Output 33.983.105
QUALITY CLIPPING
Quality cut-off: 20 MIN % 90 961.635
OUTPUT 33.021.470
MERGED READS 5.726.244