duplicated RNA-Seq libraries

[mjp]

We just had our first sequencing run performed on Hi-Seq.
We have done RNA-Seq experiment and it turned out that our libraries were heavily duplicated. The level of duplications varied from 35% up to 90%!

Somebody mentioned that can be due to library saturation when working with a small transcriptome or it can be PCR overamplification.

I don't think it's due to PCR but rather due to relatively small transcriptome we are dealing with (Arabidopsis) - but still 90%?

Have you seen such high level of duplications before? Is there any alternative reason for that?

The other question is: if this is due to library saturation how can we avoid that in future runs?

[A Oshlack]

I am no expert on this but I have heard that if you set your size selection window too small you get this sort of problem.

[mjp]

The size selection was not too small. In fact it was 400-600bp and it worked well before. Can't blame that then.

[mgogol]

If it's due to library saturation you could multiplex samples to save money.

[mjp]

We have actually multiplexed them. We used two lanes with four samples in each lane.