Quantitation Variations with Qubit

[tmm343]

Hello All,

I have run into quite a strange issue recently.

I have successfully and consistently been running 16S and 18S libraries and quantifying the final multiplexed pool using the Qubit. Recently I used a different assay then normal on the Qubit (BR instead of HS) and had a failed sequencing run. After going back and re-quantifying I saw that the BR and HS were yielding two different concentrations and always a twofold difference. BR always 2x higher than the HS reading. From our sequencing results, it also seems as the HS is accurate for the 18S sample while the BR is accurate for the 16S.

Has anyone encountered this before or have any idea as to what could cause this?

Thank you!

[nucacidhunter]

It is most likely due to concentration of library. HS assay is recommended for library quantification because of its accuracy in low range.

[microgirl123]

I don't get the same results from the BR and HS assays either. I think the HS assay is just not correct at the top end of the standard curve. I usually pool samples based on the HS assay and then use qPCR to quantify the pooled library - my pooled library is often twice as concentrated as it should be based on the Qubit.

Was the quantification of the pooled 18s library very different from the 16s library or were they both in the middle of the standard curve? And you are converting the values to nM using the size (in bp) of the library?

[nucacidhunter]

I don't get the same results from the BR and HS assays either. I think the HS assay is just not correct at the top end of the standard curve. I usually pool samples based on the HS assay and then use qPCR to quantify the pooled library - my pooled library is often twice as concentrated as it should be based on the Qubit.

Qubit and qPCR quantify different properties of library. Qubit quantifies dsDNA regardless of adapter presence in fragment termini while qPCR quantifies the fragments flanked by adapters that are amplifiable. The quantity measured by these methods will be closer to each other if all fragments were double stranded and flanked by adapter sequences.

Qubit will be fine for amplicon libraries quantification where the PCR input was low so there is not much carry over of non-adapted DNA.

[MU Core]

Consistency can be an issue. Our group performs two independent HS Qubit readings per library. Doing so will help catch any outlier measurements due to a pipetting error, etc. The average is then used between the two readings for nM calculation. Using the average fragment size has been shown in our hands to be more consistent as well. Although this is less of an issue for amplicons.

[tmm343]

Hi All,

We normally qPCR but for these 16S and 18S libraries we have found Qubit to be more than sufficient and have been using it successfully and consistently and just recently ran into this issue and seeing this discrepancy.

How we came across the issue was using the Quanit BR reagents and a plate reader for the individual libraries and doing a final QC with the Qubit on the pooled sample as a double check. The Qubit assay initially used was the HS. While the concentration and pooling was done with the Quantit BR to a 4 nM pool the Qubit HS told us the pool was at 2 nM. We decided to use the 2 nM concentration and overclustered our run leading us to believe the BR assay was initially correct. We decided to go back and do this comparison and found the variation to be between the BR and HS assays which is showing using both Qubit and Quantit reagents. It becomes even more complicated the for 16S and 18S samples we see this discrepancy but seems as if the HS is accurate for the 18S samples and BR is accurate for the 16S samples.

@microgirl123 we are using fragment size to do the conversion to nM. The sizes are pretty consistent between all the samples.

[DPVT]

qPCR/ Qubit discrepency

Our lab has often encountered trouble quantifying our Illumina pools. We sequence all manner of samples and find that some pools cluster more robustly then other others even when loaded at the same molarity. As a work around we load pools at different concentrations according how different the qPCR concentration and qubit HS concentration is.

[chuanwu]

Hi,

Did you believe the concentrations according to the qPCR or the qubit HS concentration?
Thanks

[MU Core]

chanwu,

We trust both measurements but use each for different library types. qPCR is used for PCR-free library preps with all other libraries which undergo amplification being measured by Qubit. The key is to find the method that works best for you and will allow you to normalize your loading of libraries. For example, we measure all RNA-Seq library preps by Qubit and know that our target loading density is consistent on the NextSeq with 0.9 pM. If you vary the method, however, you'll definitely see loading variability.

[DPVT]

Hey Chuanwu and MU Core,

We've started to include KAPA standard 0 (from their illumina quantitation kit) to our qPCR assays and have adjusted the qPCR concentrations of our pools accordingly. In doing so we've had much more consistent cluster densities on our MiSeq/HiSeq runs.

[chuanwu]

Hi DPVT,
Thank you for your reply. How did you do 10,000 or 20,000 times dilution of your samples? Did you try two consecutive 1/100 dilutions or more than 2 consecutive dilutions? Such as 50X, 200X, 2000X, 10000X etc?

By the way, have you ever measured your pooled library again by qPCR, whether it really matches?

Thanks

[adam.geber]

DPVT,

Can you elaborate on how you adjust your qPCR measurements? Do you simply scale everything according to the discrepancy between the expected and measured molarity of the internal 200pM standard?

Thank you!

[lac302]

I'm not a big fan of the 2pt standard curve used on the Qubit. Always gave inconsistent readings. We've actually stopped using it and rely solely on NanoDrop readings for amplified libraries and qPCR for PCR-Free libraries. Never had a problem on a hundred or so MiSeq and NextSeq runs.

[DPVT]

So for our qPCR we do 20,000x dilutions of our pools to achieve a dilution in the middle of the stnd curve (around 0.2pM) and do two dilutions of 1000x and 5000x of the internal standard 0. however much the average of those two standards is off from 200pM is how much we normalize the calculated concentration of our pools by