IonTorrent pre-miRNA data

[miRNAworld.com]

Hi am newbie so be kind,

Been trying to analyse iontorrent PGM data of low complexity pre-miRNA sequence population (less than 100 different miRNAs) encoded in lenti-viral vector. fastq outputs about typically 50k reads per sample (barcoded). Been working mainly in Galaxy because have limited bioinformatics skills so been a rough ride. Big problem is read length varies from 20-200 mer with average 140 (don´t forget these are pre-miRNAs). Removed vector sequences best I could tried several miRNA-specific approches and found best to be Dario server but problem is with input max length which is always small (45nt max). As the mature sequence can be anywhere (5´3´or middle) in pre-miRNA seq it is imposible to trim to right size.
So have used bowtie against hg19 and then converted to bam and used cufflinks then cuffcompare against hsa.gff reference dowloaded from mirbase. I am after two things reasurance that this is a good/best approach and secondly a better way than FKPM to look at data as I don´t think this type of data is it particularly suitable.

Thanks!!!!