I would like to do a qPCR on multiple samples, more than what a single plate can handle. These samples are coming from 4 different sources and represent a kinetic after stimulation with a reagent. I would like to test two genes, three if we count the house-keeping gene. In my analysis, I'm checking at the end the balance between my two measured genes to have an idea of the state of my cells : Gene 1 divided by Gene 2, the result gives me my answer.
What is the best sample repartition in the two plates in order to be sure that everything is comparable ? Should I :
1 - Put all samples in one plate, meaning doing two plates, each of them with the house-keeping gene for all the samples and measuring only one gene of interest by plate. It seems better to me but it uses a lot of reagents.
2 - Put only half of the samples to ensure to check the two genes for each sample at the same time and put the remaining samples in the second plate, while ensuring that I reput the unstimulated reference sample in the second plate to be able to do the Delta Ct in each well separately. I would also reput the blank in the second plate.
3 - Do just like 2, but not reputting the unstimulated reference sample or the blank in the second plate.
I precise that the two plates will be done at the same time with the same mixes.
Thank you for your help.
If you are interested in the ratio of Gene1:Gene2, why do you measure an additional gene ("house-keering gene")?
Note that the ratio obtained from the ct-values will be kind of arbitrary (it is an apparent ratio, not a ratio of actual molecule concentrations!) and interpretable only in a comparison to other ratios (calculated in the same way from the same kind of data). For instance, getting ct-values of 20 and 23 for gene 1 and 2, respectively, would give you a dct (log2 ratio) of (-20)-(-23) = 3 (gene1:gene2). This value does NOT tell you that the concentration of the gene1 target sequence was 2³ = 8 times as high as the concentration of gene2 target sequence! If these ct values were from a control sample, then you can compare the delta-ct to that of a expermental (treated, whatever) sample. If the ct-values from the experimental sample were 19 and 21, giving a dct of 2. These two dct values can be compared: ddct = 3 - 2 = 1, what is the difference of the log2 ratios in the two conditions.
If the ratio should give you the "real" concentration ratio of the target sequences, then you will need to perfrom a quantification relative to an external standard with known concentrations of the target sequencs (->calibration curves).
If the relative differences between conditions are sufficient, you should put the reactions to calculate the dc values from on the same plate. This is possible for any number of samples (it will not work in the unlikely case that you want to run more than 48 technical replicates per gene and sample ;) ).
If you want an absolute comparison fo the target sequence concentrations within a sample, then you will need to measure good calibration curves. You can run the two curves (one per gene) on each plate, or you can measure them once one a single plate, and further including only one standard dilution on each plate with samples. The ct-values from each "sample-plate" must then be corrected for the difference of the ct-values of the used standard dilution to the corresponding dilution of the "calibration-plate".
I'm sorry, it looks like I wasn't clear in my message, I apologize. I will reformulate and add precisions :
I am not comparing directly the ct values, I first calculate the dCT thanks to my house keeping gene, then I calculate the ddCT by calculating Stimulated-Unstimulated to first have an idea of the increase or decrease that I have for each gene. This result is the first that I look, the most important one.
In fact, the variations of Gene 1 are pretty clear and readable and I can draw clear conclusions from them, which is not the case for Gene 2 for which the variation are very small. However, Gene 2 is highly linked to Gene 1 and I needed a tool to see this influence with a qPCR, that's how Gene1/Gene2 ratio came out as the Gene 2 variations are between 0 and 1,5 ddCT.
In the end, my last calculation, Gene1 / Gene2 is a ratio done on the DDct which is completely arbitrary for now, I talked about it just to give more precisions on my work. My major concern was for the comparison of the timepoints.
Indeed, I could see from the work of a collegue who works on the same cells than me that he didn't have the same House-keeping gene CT value in two different plates for the same sample. For his qPCR plates, he uses the second method that I describe in my first post. From there, I was concerned about my results, if do like him, half of my timepoints are separated between two plates and I was not sure that they would be comparable anymore if the house-keeping gene is already so different between two plates using the same mixes (difference was more than 2 CT from the same sample). In consequences I prefered until now to do the first method by separating the genes in two plates, but not the sample to be sure that for one gene, every timepoint is still comparable with the other one, but at the cost of a large quantity of reagents due to the fact that I redo the house-keeping gene measurement on each plate.
In the end, my question is which method is the best to be sure that everything is reproducible and scientifically valid for the measuring and comparaison of the timepoints, my ratio Gene 1 / Gene 2 is only secondary and is just here to help me to draw conclusions from Gene 2 variations.
Thank you very much for your answers.
You are on the safe side when you calculate dCt values only "within plate".
- Badges
- Science method