different duplication levels in forward reverse libraries

[Mchicken]

Hi guys,
we are currently analyzing RNA-seq libraries from tomato and used FastQC for this. Remarkably, comparing the duplication levels from the forward and reverse library of the same experiment we see a different pattern in duplication level.
The plots you can find attached show, that the reverse reads are a lot more duplicated than the forward reads.
Do any of you have an explanation for this phenomenom? We expected a similar pattern for forward and reverse library.
Thanks a lot!

[nucacidhunter]

Would you give some information regarding the library type such as stranded, non-stranded, total RNA or mRNA .

[Mchicken]

It`s a strand specific and polyA-purified library with 150bp long paired-end reads. So we should have only mRNAs.

[nucacidhunter]

I assume the plots are for raw data with no adapter trimming. Also it is sometimes helpful to have other FastQC plots such as for adapters, overrepresented sequences and quality.