RNAseq Log2Fold Change and the 2-Delta Ct fold change

Question

[Jonathan Crowther]

Hi All,

I just have a quick question/confusion.

I have performed DEseq analysis of some RNAseq data and found genes to be quite nicely deferentially regulated.

Another fellow in the lab has performed the qrt-PCR and we have found our data to be quite similar. I am curious about the 2-DeltaDelta Ct normalization. I see that our fold change from qRt-PCR match our Log2Fold change data from Deseq and am curious as to what point in the qrt-PCR has the log2 Transformation as I cannot see it directly. I am assuming the 2** step is most likely the case but this has me confused as I would expect the 2** step would take the data from a log2 and hence should make the comparisons to the Fold change in the DEseq data not the Log2 Fold Change.

Any help would be appreciated.

Answer 1

[Charles Warden]

Ct values are already on a log2 scale: they represent the number of cycles required to observe a certain threshold of PCR sequences, which doubles each cycle (1 cycle = 2x original sequence abundance, 2 cycles = 4x original sequence abundance, etc.). Therefore, higher Ct values correspond to lower expresion levels (because the original abundance was lower).

[Jonathan Crowther]

Hi Charles

Great that makes a lot of sense, but then does the 2**(-deltadeltaCt) mean it takes it from the log2 Space to represent the Fold Change or does it Stay in the Log2FoldChange space?

My impression is the 2** will remove it from the log2 space.

Cheers

[Charles Warden]

"fold-change" and "deltadeltaCt" can vary a bit their definitions. The transformation you describe will convert from a log2 to a linear scale. Whether that is the precisely right strategy depends upon what you want.

For example, my guess is that the first deltaCt was a normalization based upon GAPDH levels: however, it matters whether Ct(GAPDH) - Ct(GeneX) or Ct(GeneX) - Ct(GAPDH) was used. Similarly, I think the second deltaCt represents the difference taken between groups (which is the part that is most similar to the fold-change), but obviously it matters which group was used as the reference.