Library quantitation discrepancy

[genlyai]

I have a strange result, and I'd love it if anyone has any hypotheses for me.

I want to know the complexity of my Illumina TruSeq library (the absolute number of amplifiable molecules prior to the pre-PCR step). I applied Kapa Biosystems' methodology (Sybr Green qPCR with supplied standard curve) to determine this. The standard curve looks perfect and all the samples fall within the standard curve.

However, just for kicks, before putting samples into the qPCR I first diluted them into Illumina's pre-PCR reaction buffer and cycled them for either 0, 2, or 5 cycles. Using this design, i should be able to measure the # amplifiable molecules present at each cycle number, and the expectation was that there would be <=4X as many at 2 cycles than 0 cycles, and <=8X as many at 5 cycles as 2 cycles. This would verify that the two types of PCR reactions were amplifying the same subset of molecules.

In my actual qPCR results, 5cyc vs 2cyc looks exactly as expected (they are exactly 3Ct apart), but there are apparently 16X as many amplifiable molecules present at 2cyc as there are at 0cyc (consistently 4Ct apart across all libraries).

Can anyone explain this? My only hypothesis is that the Illumina denaturation (98C for 30s) makes more molecules available for amplification than the Kapa denaturation (95C for 5min), but this seems doubtful. How would you determine the true starting library complexity as amplifiable by Illumina's kit: just take the result from 0cyc, or back-calculate from the 2cyc and 5cyc results?

Apologies for the long post. To those who got through it, thanks for any help!

[Eric@Kapa]

These ideas are just that -- purely spur-of-the-moment hypotheses .

Perhaps unamplified ("cycle 0") libraries amplify relatively inefficiently (initially) because:

1. There is still ligase hanging around on the DNA, inhibiting DNA denaturation and/or pol extension. How did you clean up your library material after adaptor ligation? Did you size-select via agarose gel?

2. DNA was exposed to UV during gel purification and was damaged -- this would inhibit amplification.

3. The phosphorothioate bond at the 3' end of the Y-adaptor inhibits polymerase, so the "original" library molecules are inefficiently amplified, while subsequent copies that do not contain pt bonds are efficiently amplified.

4. Adaptor ligation is incomplete (either adaptors aren't fully phosphorylated, or PNK treatment during end-repair is imperfect, leading to only one of the two strands being ligated in a subset of cases), and some "nick translation" as per Nextera must occur to complete the Y-adaptor. Of course, if this were true, then it would require some kind of PCR-mediated "recombination" to get the adaptor (or at least primer) region onto the end of the incomplete strand.