Hello,
I'm an undergrad and I'm trying to figure out how to deal with qPCR data.
I'm currently working on this :
I have 7 experimental conditions (DMSO for control, 3 concentrations of a drug A, and 3 concentrations of a drug B), in which I treat human cells.
For each treatment, I did like this : I took 3 wells of cells from the same thawing, then I treated each well with the same treatment.
Then I lysed all cells and instead of pooling each lysat together like I usually do, I made one well = one Eppendorf tube. So I ended up with 3*7=21 eppendorf tubes, which I annoted DMSO_1, DMSO_2, DMSO_3 for DMSO samples well 1, well 2 and well 3, etc.
Then, I extracted the RNA from each sample, one third at a time (every "well 1" samples together, etc.), not for any particular reason, I just didn't feel comfortable extracting the RNA from all my samples at once.
Then I made my RT one third at a time again, and then my qPCRs, using 2 different reference genes (one for drug A, one for drug B).
So from start to finish, all my "Well X" samples are grouped together and separated from my other "thirds" of wells.
At first I thought I would just end up pooling all the relative expressions for each genes and each samples. But then my supervisor pointed out that my DMSO samples may have been different from one another, and that maybe putting DMSO1 with DRUGA_1 wasn't the right thing to do; and therefore I needed to find out how to analyze all the data from separate RT.
But I honestly don't know how to deal with that. The easiest solution would be to perform RT-PCR all over again, by doing the RT for all my samples at once, and then put everything on a same qPCR plate, but I would like to avoid doing that to save some money. Instead, I would like to know if there is some way to use all the raw data obtained for all my different RT-qPCRs.
Would making the average of all my raw expression data from each condition then normalizing on the averaged DMSOs work? (Using the deltaCT method)
I hope this isn't too confusing (or cringy to read, because maybe not doing that in the first place would be obvious to you).
Regards
Very good question!
If I understand you correctly, you got "bad" results (i.e. unexpected or inexplicable) and now want to know what went wrong, right? And your supervisor pointed out that "weeell, maybe your RT procedure is to blame", is that right?
First of all, I think you need not worry for RT (if I read you right, it's Reverse Transcription). If you keep the procedure same, your cDNA will consistently yield same results, even if you treat all your samples on different days. When you do the RT absolutely does not matter as long as you store your resulting cDNA under right conditions (freeze).
Second, your supervisor is wrong in that you CAN and actually SHOULD compare your DMSO samples to drug-treated samples individually. It is not a good argument that "maybe your DMSO samples were different from one another". If they are, in fact, different, it means that DMSO or the procedure is somehow affecting your results, and you HAVE TO find that out and make sure everything goes well! You cannot dismiss such finding, that your control samples are vastly different from each other, it is wrong scientifically.
However, I suspect that your problem is not the difference of DMSO samples.
If I read you right, you ran all your samples on different qPCR plates, is that right? All your "Well X" samples were separated from other thirds in all steps, including qPCR? If that is right, I think your biggest problem is not running all samples on 1 qPCR plate at once. See, the "threshold line", by which you obtain Ct (or Cq) values is drawn by the machine FOR EACH PLATE individually. Moreover, conditions of heating are different between different plates that you run and result in differences in fluorescence readings. All of this means that you absolutely have to run ALL samples you intend to compare on the same plate, or you will get different results.
If you ran several plates, your data will not be comparable, and there is probably no way to fix it, unless you made sure that the starting DNA concentration was same in all samples. Did you ensure the cDNA concentration was same or did you just run without checking your DNA concentrations (right after RT)?
If you showed me your raw data, I could be able to think of some solution. Please feel free to contact me directly, I am interested in your problem (I investigate problems in modern qPCR protocols and what kind of problems people usually have and how to fix them).
One more thing you may want to check is your melting curves. Are they okay? If you post pictures of melting curves here, I could instantly identify the problem (that happens too).
Anyway, please feel free to write me.
Yulia gave a great answer.
One thing I see differently: there is no absolute need to have all samples under all conditions on the same plate. For larger experiments, that's not even possible. But then, the amplification curve data should be combined from the different plates, and a common baseline correction and threshold should be used for all curves from the same gene (across all conditions). Only then you get reliable dCt values. Not all software that is shipped with qPCR instruments is able to do this, in which case you would need to do this "manually" (I don't know a neat software that would do this, but there are packages available for R that should do it: "qpcR" and "EasyqpcR", maybe others; however, if you don't know R that's quite challenging).
Was there a particular reason why a different reference gene was used for the two treatments?
Jochen Wilhelm Thank you for your answer. Indeed, at first I used the same gene for both drugs, but it appeared after a few tries that my reference gene had different expression levels at high concentration of drug A, while the second reference gene I tried had different expression levels at high concentration of drug B. I actually thought first that my RT was to blame for this, but this happened everytime on each of my "thirds" with the same condition, so I just assumed I should switch reference gene for each drug. I indeed don't know R so that might be a problem. The systeme we use has this software "CFX manager", which includes a "gene study" function, but my supervisor has told me that he never uses it to analyze data obtained from different RTs.
I have downladed R and the package "easyqpcr", it seems to be designed to treat such case, however it requires a "calibrator" sample, which I do not have. If I get it right, the calibrator is a sample for which I know the exact concentration. Then, can I mesure my single strand DNA concentration in my control samples, and use them as calibrators?
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA