Help understanding FASTQC output

[gkandoi]

I'm using this data for RNA-Seq analysis and when I run FastQC on it, I get some weird graphs for the Kmer Content and Sequence Duplication levels.

Can someone help me understand what could've possibly caused this? I ran TopHat on the data and it says ~94% mapped, which I think is fine enough.

[cmbetts]

The FastQC flagging parameters are generally tailored toward WGS data where an ideal result would be uniform coverage.
I don't think I've ever seen RNA-Seq data not fail the kmer or duplication level checks. They both fail due to a combination of the non-uniform abundance of all of the different RNA species and the fact that the "random primers" used for first strand synthesis show a decent bit of sequence/GC bias in their priming.