10 replies to this topic

[Podge]

Hey guys,

I'm running RT-qPCR for relative expression on tissues.

I ran some testsamples on with conventional PCR (cDNA 1:10) and looked at them on a gel. In comparison to my reference gene the bands look alright and I expected them to give me nice Cqs. Also used 4x the cDNA in the testrun, but the bands on the gel actually didn't look different.

Ran the lot of my samples in the qPCR then, but only get Cqs of 37-39.

I am using the Roche UPL system, so the amplicon is detected by a probe that is specific.

Any ideas what could be the problem?

Thanks!

[bob1]

PCR on a gel is an end point system - if you have enough DNA you will be able to detect it, the brightness of the bands is only partially related to the efficiency of the PCR.  Even a few cycles of a PCR will amplify the DNA enough that you can see it on a gel.

It could also be that you have some sort of inhibitors in the cDNA (carried over from extraction) that are inhibiting the PCR itself, these might be diluted out sufficiently in a 1:10 dilution such that the PCR works better.

[Trof]

Your real-time reaction have  different conditions and probably reaction composition.
Did you looked at the gel of the product from the real-time reaction?

And actually in UPL system probes are not specific, only primers are.

[Podge]

Thanks for your answers, guys!

bob1, on 12 November 2012 - 11:21 AM, said:

PCR on a gel is an end point system - if you have enough DNA you will be able to detect it, the brightness of the bands is only partially related to the efficiency of the PCR.  Even a few cycles of a PCR will amplify the DNA enough that you can see it on a gel.

I see. Good to know. Thanks

Quote

It could also be that you have some sort of inhibitors in the cDNA (carried over from extraction) that are inhibiting the PCR itself, these might be diluted out sufficiently in a 1:10 dilution such that the PCR works better.

Nanodrop values are superb. 260/280 is 2.15 in average and 260/230 is 2.27 in average for my samples
I'm using RNeasy Lipid extraction kit, which uses Trizol in combination with spin columns.
I elute RNA in 60ul H20. Plenty of good quality RNA there.

Quote

Your real-time reaction have  different conditions and probably reaction composition.
Did you looked at the gel of the product from the real-time reaction?

And actually in UPL system probes are not specific, only primers are

I use the same chemicals for conventional (minus the probe) and real-time PCR. Only the thermocycler and tubes/plates are different. I also put some of my qPCR samples onto the gel, which shows a band just like the conventional PCR.

Edited by Podge, 13 November 2012 - 02:43 AM.

[Trof]

So, you get bright bands on gel, but low and/or late florescence? Then the probe is the problem. Did you design the primers using Roche Design Center? Did you took the best primers?
Question is, whether UPL system works in your conditions at all (maybe noncompatible mastermix or something) or whether this particular assay is bad. Because they sometimes can be.
Are you using Roche mastermix, cycler?
Did you had ever a good working UPL assay or is this your first? If you did, then I would just design new primers/assay with a different probe if you have probe set.
If it's your first time (or you have several assays that have same problem), maybe I would try to optimise, but the only thing that you really can if your reaction runs fine is probably concentration..
Also check you are using the right channel (FAM).

[bob1]

Podge, on 13 November 2012 - 02:01 AM, said:

Quote

It could also be that you have some sort of inhibitors in the cDNA (carried over from extraction) that are inhibiting the PCR itself, these might be diluted out sufficiently in a 1:10 dilution such that the PCR works better.

Nanodrop values are superb. 260/280 is 2.15 in average and 260/230 is 2.27 in average for my samples
I'm using RNeasy Lipid extraction kit, which uses Trizol in combination with spin columns.
I elute RNA in 60ul H20. Plenty of good quality RNA there.

Trust not only in Nanodrop - run some on a gel too to check for integrity of the RNA.  Note that 260:280 ratio of 2.15 is high for pure RNA, value should be 1.9--2.0.  What did the curve look like?  If there was a peak or shoulder in the 230 range then you probably have phenolic contamination or potentially some protein carry over.

[Podge]

Trof, on 13 November 2012 - 09:19 AM, said:

So, you get bright bands on gel, but low and/or late florescence? Then the probe is the problem. Did you design the primers using Roche Design Center? Did you took the best primers?
Question is, whether UPL system works in your conditions at all (maybe noncompatible mastermix or something) or whether this particular assay is bad. Because they sometimes can be.
Are you using Roche mastermix, cycler?
Did you had ever a good working UPL assay or is this your first? If you did, then I would just design new primers/assay with a different probe if you have probe set.
If it's your first time (or you have several assays that have same problem), maybe I would try to optimise, but the only thing that you really can if your reaction runs fine is probably concentration..
Also check you are using the right channel (FAM).

Hey.
Yes. Band on gel, but no signal/very very late signal in qPCR.
I always design the primers using the design centre and use one of the top5 that matches all prerequesites. Saying that I have ran many a UPL assay. So it is not the problem of mastermix or the real-time machine. I do you corresponding reference gene values for my samples, so it has to be in the primers or probe.

By "If it's your first time (or you have several assays that have same problem), maybe I would try to optimise,..." you mean optimising primer concentration etc.

I have now designed new primers and ordered them.

Quote

Trust not only in Nanodrop - run some on a gel too to check for integrity of the RNA.  Note that 260:280 ratio of 2.15 is high for pure RNA, value should be 1.9--2.0.  What did the curve look like?  If there was a peak or shoulder in the 230 range then you probably have phenolic contamination or potentially some protein carry over.

Good to know the value should be between 1.9-2.0. I somehow always thought a perfect 260:280 ratio is 2.3!
Here a screenshot of my 16 samples. Apart from one with a crappy value (light blue), all of them were good.
I will run some RNA samples on a gel later. Do you know how much one adds to a gel normally?

[bob1]

Your curves look pretty good, not much absorbance in the 230 range, which indicates that there isn't much contamination with phenolics and proteins.  You need to run about 200-500 ng of RNA on the gel to be able to see it effectively.  YOu should be able to see a minimum of 2 bright bands (18s and 28s), possibly a few others (also sRNA) and maybe a smear in the lower part of the gel, which, if the sRNA bands are present, is usually mRNA.  If you only see a smear - degradation is your problem.

Degraded RNA and DNA will still give you a reading on the nano-drop, even nucleotides will!

[Trof]

bob1: I think if there was something wrong with the PCR itself, he wouldn't get good bands on gel from the same PCR product that has high Ct. The probe is almost surely the problem.

Podge: Optimising.. well, primers maybe, but I'm not sure what effect it would have on the binding of the probe, by running the product on a gel you proved your PCR is running fine, so maybe optimising probe concentration or something (or maybe even the primers/probes ratio, since there may be some interaction). They claim I think 96% success rate with one of first three primers designed by Design Center, so sometimes you fall into that 4 %.

In case a UPL design fails I order another, if that's not possible for some reason I just use the primers, which are good and run the assay on SYBR Green. UPL is not designed to be optimised much.

[bob1]

Trof, on 14 November 2012 - 02:01 PM, said:

bob1: I think if there was something wrong with the PCR itself, he wouldn't get good bands on gel from the same PCR product that has high Ct. The probe is almost surely the problem.

Yeah I agree, having seen your reasoning, I just thought from the initial post that something could be wrong with the RNA, and that carried on for a bit.  Definitely not my area of expertise, so I'll leave it to you.

[Podge]

Hey guys,
sorry for the late answer, but was on some holidays for an extended weekend.

Here are my results with this particular PCR.

UPL:
cDNA 1:2.5: Cq ~35
cDNA 1:10: Cq ~37-39
Gel: right band at 132bp

SybrGreen:
cDNA 1:2.5: Cq ~32
cDNA 1:10: Cq ~28
Gel:     primer dimers appear
    Product not the right size (200bp instead of 132bp)

• Different size for 1:10 and 1:2.5!!

Melting curve:  Primer dimers visible!

    Slightly different Tm for 1:10 and 1:2.5 samples

All in all it seems like the whole Primer plus Probe combination is messed up. Dimers. different band sizes?! Got new primers today and will test them.
At least I know the gene is there in some way. Tried optimising this PCR in a different tissue where we weren't sure about expression levels. At least I know now it was the assay itself.