I have done qPCR with 3 groups: 1 control, and 2 drug conditions, this qPCR only targeted one gene without being done with a reference gene. My questions is:
If I detect a change in the log 2 values between my controls and conditions, is it acceptable to claim that there is a difference between the groups in terms of the GOI that was measured, even without normalization to a reference gene?
Example: when compared to control, the average log2 change in one of the treated groups is (-0.41). Is this data meaningless without a reference gene?
Thank you for any advice.
Hi Ricardo Peguero
just imagine that your PCRs do not follow the same specificity and efficiency, how to compare them?
the best thing is to follow the ddCT protocol to be sure to avoid lot of biases.
see attached an xls file for calculating the FC between PCR.
all the best
fred
Frederic Lepretre Thank you very much for the resource. As far as the qPCR data I produced without using a HKG - are these data irrelevant? I would believe so as you mentioned potential differences in specificity/efficiency
You have to include a housekeeping gene, that's the only way to compare between different biological samples.
Your data are not interpretable or publishable without the full set of controls.
If you have "difficult" sample material (tissue or organ samples), it is extremely difficult to have comparable samples (with comparable RNA yield, integrity, RT efficiency). Using a reference gene is a convenient way to control for all these sources of differnces. However, with considerable effort you may control all these factors very carefully and so end up with comparable Ct values. But this is difficult and reqires a lot of work and reporting. For some "simple" sample material (well-defines cell cultures), one might assume that integrity and RT efficiencies are comparable between samples, and so it may suffice to take care that the same amount of starting material is used in the qPCR. If this is publishable depends if you can convince the reviewers that you did all controls and that your work-flow is robust.
And another factor is the amount of differential expression you want to discuss. If this is huge, you will robustly detect it with or without proper normalization (but in such cases qPCR is actually like using a sledgehammer to crack a nut). If it is not that huge, normalization may lead to the opposite conclusion (what is not a nice situation, as you may imagine).
I guess the real question is "why have you not used a reference gene?"
There may be scenarios where a reference gene is impossible or simply doesn't make sense: if you were, say, detecting bacterial genomes in groundwater, where there is no such thing as an "endogenous control" (coz it's groundwater), then you might have to resort to either simply stating that all effort was made to standardise extractions.
Alternatively, you might consider using a spike-in control (i.e. adding some easily detected gene marker to your groundwater before DNA extraction, as an internal calibrator).
data without HK controls just give you an idea on what's up, but just an idea, no way to compare samples or conditions. A reference is always needed in PCR studies.
which leads the importance of preparing good experimental protocols before going to the bench.
all the best
fred
Jochen Wilhelm John Hildyard Frederic Lepretre thank you for your answers. As my PI put it to me, much like what you are saying, Frederic, I conducted this assay with no HKG simply to "have an idea of what's up" before conducting more robust analyses. Thank you all again for your time.
Hi Ricardo
You might find this paper interesting
Article No control genes required: Bayesian analysis of qRT-PCR data
- Badges
- Science method