Illumina RNA TruSeq kits and "bottomed-out" libraries

[pmiguel]

What final library concentration do you consider "good" for an Illumina RNA TruSeq library? 10nM?

If you presume the 15 cycles of amplification recommended by the kit produces 1.8x increase of amplicons per cycle, then I think your initial number of amplicons (unamped library molecules) could be as low as 9 million. At least that is what I get when I run the calculations. (Presuming the final amped libraries is eluted in 10ul.)

That may result in an acceptably low amount of PCR duplicate in a 10 million read data set, but not so great for 30 million.

Anyone else surprised by this?

(Please check my math, if you are so inclined.)

Comments?

--
Phillip

[kmcarr]

I'm going to bump Phillip's question because it's a good one and nobody answered

Also I'd like to expand on it with these questions:

What is the average final concentration you get for your TruSeq mRNA Seq libraries, assuming you started w/ ≥ 1.0 µg of Total RNA input? Also, do you do any sort of final size selection after the final enrichment PCR (e.g. SPRI cuts)?

Same question for TruSeq DNA libraries, again assuming you started with 1.0 µg of DNA? Do you size select your DNA libraries and if so how?

Thanks.

[Sciurus]

I'd be interested in that too! Anybody any idea? :-)

[HESmith]

Since no one else has responded to Phillip's question, I'll take a crack at it.

The mistaken assumption is that the efficiency of amplification is ~2 (1.8 in the original post) for every cycle. The reaction begins to approach saturation well before the end of amplification (think of those S-shaped curves you get from qPCR), so the latter cycles increase the amount of product <2X. For example, a 5 pM starting concentration of template will be past the linear amplification range by cycle 15. However, you still get some amplification (>1X) past the linear range, so the latter cycles serve to increase the total amount of product.

If you've made libraries from different amounts of input, you've already observed the saturation effect in action: 1 ug of DNA does not yield 10X as much library as 100 ng.

You can also determine from your data that 30 M reads is not merely 3X PCR duplicates of a 10 M molecule library, simply by examining the degree of read duplication after alignment. Until the read depth begins to approach the gene length, the majority of reads will be unique (i.e., align to different positions).

HTH,
Harold