qPCR for telomere length measurement - efficiency issues - complicated monochrome multiplex assay with SYBR Green (Aug/17/2010 )

Hello,

I am trying to get a qPCR assay to measure human telomere lengths up and running (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2647324/). It uses SYBR green chemistry on an iCycler iQ (BioRad) machine and is multiplex (telomere and control single-copy gene amplified in same tube) by using some fancy primer design. My efficiency values are tending to be too low (often <80%). I think the first advice most people would give me is to redesign my primers, but given some complex primer design and that I am trying to replicate the published protocol which others have used successfully, I am hoping to stick with these primers. What other things would you recommend first to try to fine tune to increase my efficiency values? I have tried increasing and decreasing my annealing temperature a few degrees without much luck.

There are some other complexities to the assay which I am trying to spare you, but please either check the above link, or let me know if there is any more information I can provide to help you help me.

Thank you very much for your time.

Dan

As a follow up, I have been told by a few others in my building that worrying about efficiencies (as calculated from standard curves) is not needed as long as my R2 is good and the traces and melt curves look ok. I've been told that the other labs who claim consistently good (90-110%) efficiencies across runs are probably exaggerating.
How much should I be worrying about efficiency values?

Hello!

I'm also trying to get Cawthon's protocol running in our lab. I have the problem that my efficiencies are too high. Did you recognized primer dimer? At the telomere Tm my NTC are coming, but far enough from the rest to worry about.

The beta Globulin Primer are not so good. Those for Albumin as SCG are better and R. Cawthon published some new Primer in a blog in this Forum (somewhere). If you want I can give those primers tp you.

Which concentration of SCG and telomere primers do you use?

First of all the efficiencies of telg/telc an SCG should be the same so you are able to calculate your T/S! Normally I find 80-90% is a good efficiency. 100% does not reflect reality. efficiency over 100% stands for contamination or primer dimer in the reaction. I would use result until 105% efficiency. After all a non-varying efficiency is the best.

Good luck, Nathalie

Hi Nathalie (and others),

I am very pleased to have someone else who understands the plight I am in

I was previously fooling around with using the Fermentas SYBR green with Fluorescein MasterMix (on Bio-Rad iCycler iQ). With that I tended to get high efficiency values as well (~110-140, but also rather inconsistent and sometimes dropping down into the 80s). Are you using Cawthon's original homewbrew mastermix recipe or an off the shelf mastermix? I assume you are trying with the 2009 protocol right?

I've avoided using the beta-globin primers too--mainly because there are several unconfirmed SNPs overlapping with those primers. Glad to hear I am at least using the better primer set on more empirical grounds. I'll look for thew new Cawthon primers posted on the blog and post them here if I find them (if not I would of course appreciate if you can post).

Using the originally recommended 900nm for all primers, I found that the telomere peak on melting curve tends to be almost non-existant on the highest concentrations on my standard curve (88ng genomic DNA). I have switched to 500 for telomere primers and 300 for albumin and that has corrected the melt-curve problem.

What do you make of the strategy employed in Ehrlenbach et al which seems to yield much lower CV values? If I am understanding correctly, they seem to embrace (rather than battle) the low efficiency values and use them to help calculate the T and S values with LinRegPCR. I am experimenting with LinRegPCR myself now. So far it looks like mean efficiency of 89% for telomere and 79% for albumin (SCG). This interesting because efficiencies calculated with the standard curve method almost always yield higher values for the albumin amplicon.

The more I read and talk to people, the more I am beginning to believe that the 85 or 90 to 105 or 110 rule of thumb for standard curve calculated efficiencies is not particularly based in empirical reality. My current, perhaps incorrect, synthesis of the literature is that anything under ~105% with a clean exponential phase and melt curve is ok as long as efficiency values are similar across samples in the same amplicon (and if LinRegPCR is employed, perhaps not even needing similar values across samples).

Thanks for the feedback and looking forward to more.

Dan

Nathalie Allard on Mon Aug 23 13:00:27 2010 said:

Found the other primers:
http://www.protocol-online.org/biology-forums-2/posts/14233.html

Have you had better luck with them?

I also wonder if the thermocycling profile Cawthon recommends is better for the iCycler as well...

On a related note, I have been extending my number of cycles out to 37 to better catch some of the low C amplicons. Because of the LinRegPCR suggesting I drop samples that haven't reached plateau phase, I am thinking of extending out to 40 or or 45.

D

Hello!

First I tried SybrGreen Supermix from BioRad, but the results were not reproducable (with three different SCG).

Today I gave EvaGreen a try (just both Albumin Primer as SCG, Hbgu is too bad). There I had quit good results with the Albugcr MM. EvaGreen shows much better results in the MC and the NTC don't show up. Albu MM was not good, the efficieny was too high (150%). But at all the T/S Ratio was nearly the same with both SCG. Tomorrow the same experiment again to verify the results.

Do you have the possibility to chek your results with southern blot or something else?

I also reduced the concentration of the Primer to 500nm for the telomere, 200nm for Albugcr and 300nM for Albu. For the SCG 32 cycles is ok. But you need more cycles for LinRegPCR, or?

What do you use as standard?

I didn't try the mentioned protocol for the LightCycler. I have a BioRad IQ but if I have still problems I will try it!

With LinRegPCR I have no experience. I downloaded it today to test it.

I hope the best to get it run!

Nathalie

I think I had problems with the Bio-Rad supermixes as well--although I would have to go back into my notes to see more. I think I didn't have much luck with EvaGreen either--but again would have to go dig up why (would be happy to for either if it would be helpful).

No plans to use a southern validation--although if I find time it of course would be a good idea. My impression of the literature is that it is beginning to be acceptable to publish without it.

I think the need for more cycles are in part because the low efficiencies result in higher Cqs. Also, LinRegPCR alogirthm seems to like it more. However, if I end up using the LinRegPCR method exclusively, it is probably not necessary as I won't need a standard curve with low c samples.

I've been using DNA extracted from myself as a standard. I considered using the O'Callaghan method of using an oligo as a standard--but I become worried about primer degradation, and lack of comparability between DNA extracts and oligo samples.

Best,
Dan

Nathalie Allard on Tue Aug 24 13:36:19 2010 said:

An update, using LinRegPCR, I can get T/S CVs on replicates down in the ~10% range. However, T/S values vary substantially between different DNA concentrations and across runs. See the attached for more details.

You don't happen to know how the albugcr primers are supposed to be better do you? What kind of difference do you notice between it and the original albumin primers? I figure I'll ask you before bugging Cawthon more.

Thanks,
Dan

Hello!

I run the PCR product on a 3% gel. ALbu has a product in the NTC which has the same size like the samples. Albugcr NTC is empty. I suppose that the Albugcr primer are nearly on the same place on the sequence because they have the same size on the gel.
To set this method up is not so easy. I'm not getting repducable results! I don't know why the variation between the runs is so high.

Did you try the cycling program for the LightCycler also on the BioRad? If yes, how does it work? This will be my next try for establishing this method. I will also try the selfmade mastermix in the paper. I'm sure that DTT and betaine could be very helpful in the PCR because of the GC clamp.

Attached is a gel I made!