I am checking the expression of a gene across different cell lines. I have used GAPDH as a housekeeping gene. But I cannot use the delta-delta Ct method as I don,t have a control group. How can I represent my data?
you could run a standard curve using one of your samples as the material. Then apply each value that you obtain to that curve
Ok, I think it's really important to consider the question you are actually asking here.
You want to know how expression of this gene varies between cell lines, and you assume (I hope justifiably) that GAPDH does NOT vary between these lines.
You have concluded that you cannot do this, because you do not have a control group. What, exactly, would a control group BE, here?
It's like saying "I don't know how heavy all these different items are, because I don't have a control item".
It doesn't matter. You can still tell which items are heavier than others.
Unless you are using absolute quantification, all qPCR comparisons are relative. All of them. If the normalised values for cell line 1 are double those of cell line 2 (i.e. 1 cycle lower), then cell line 1 expresses twice as much as cell line 2. No need for a control, because all the control would do is add another (spurious) data-point to this: if you had a control and it gave values half as much as cell line 2, then you could say "cell line 1 is 4-fold over control, while cell line 2 is only 2 fold", but note: cell line 1 is STILL double the expression of cell line 2, and it is the differences between cell lines that you are interested in.
So. Determine the Cq for your GOI, and the Cq for your reference gene (I would add another ref gene: always use at least two). If you're dead set on dCt methods, then subtract the ref gene Cq (or the average of the two ref gene Cqs) from the GOI Cq.
Note: Cq values are log values, so subtraction is equivalent to division in linear terms.
You now have normalised (log2) expression values for each sample. The relationship between these is what you are actually interested in, and that relationship does not care what the actual numbers are. A two-fold increase is a two-fold increase whether it's 10 vs 5, 20 vs 10 or 400000 vs 200000.
For convenience, you could average all your dCt values and subtract that from all your individual dCt values: now anything "exactly average" has a fold change of zero, and everything else is either above or below the overall average expression. Note: this has not changed the relationship between your data at all: anything that was 2x higher before will still be 2x higher now (1 cycle difference).
Or you could assess the dCt values alone.
In all of these you should be careful to make sure you're keeping track of what the numbers SHOULD be telling you, because with logs and subtractions that can flip the sign of the value multiple times, it can get quite confusing (if a sample has a GOI Cq of 22 and a ref Cq of 23, and another sample has a GOI Cq of 25 and a ref Cq of 22, the second sample SHOULD report lower expression, if you've done it right).
Or you could convert everything to relative quantities
2^(lowest Ct-sample Ct)
which linearises your data, and then just treat it like any other linear data (so for reference genes you now divide rather than subtract, etc). I like this approach, because logs are horrible.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Cell Biology
- Culture Cells
- Cell Line