Calculating Rt-PCR values using the delta delta Ct method. Should I omit the positive and negative signs?

"or do we subtract the smaller from the bigger whatever it is"

No. No no no. Never do this. You are normalizing your GOI to your reference genes, and you should ALWAYS be normalizing your GOI to your reference genes. What you are proposing is, under certain circumstances, to normalize your reference genes to your GOI. Not only is this wrong, it's actively meaningless.

One of the greatest problems with the ddCt method is that it requires you to be entirely comfortable with log transformed data, and to be fully aware at all times what each number means, and what values you are comparing with what other values. It is very easy to lose track (I prefer to use relative quantities for everything, personally: this also lets me factor in variable efficiency if needed).

If your GOI has a high Cq, that means it's low expression. If the Cq value drops in some samples (while your ref gene Cqs stay more or less the same) this corresponds to a marked increase in expression, and this is fine, even if it makes your dCt values go negative.

Sample A: GOI Cq 28, Mean Ref Cq 25 => dCt = 28-25 = 3.

Sample B: GOI Cq 20, Mean Ref Cq 24 => dCt = 20-24 = -4.

Just looking at the raw data, you can see sample B is showing massive upregulation (28->20), even accounting for the fact the ref genes show there's more cDNA in that sample (25->24), so you know that the number you get at the end should reflect this.

What's the difference between sample A and B? ddCt = 3 - -4 = 7.

And (assuming perfect efficiency) thus 2^7 = 128, so ~130-fold increase in expression in sample B, relative to sample A.

You could equally compare sample B to A: -4 - 3 = -7, 2^-7 = 0.0078, which is of course simply 1/128, i.e. sample A is ~130-fold LOWER in expression than sample B.

All your expression values will be relative, and how you present your comparisons is up to you, but just make sure you know what the numbers mean.

Again, I'd recommend switching to RQ values (efficiency(lowest Cq - sample Cq)), because this linearises all your data at the outset, and thus gets rid of confusing logs, but it's up to you.

Attached is an excel sheet showing ddCt and RQ handling of the same dataset.

What can explain a low expected heterozygosity rate for T. monococcum wild samples (0.09) ?

What is suitable temperature at reaction mixture of lipase activity assay?

What is the relationship that determines the depth of the mechanical cultivator with the age or size of the weed For crops planted in lines?

Does the subsoil have an effect on the spread of weeds whose roots are more than 35 cm deep?

Is there any evidence that using ball milling in intervals, rather than one long run, improves the mixing process?

Can I measure serum and not tissue succinate dehydrogenase?

How to monitor and mapping crisis event tweets based on my tweets dataset?

Which unsupervised learning methods can be used to detect fake users?

My melanoma cell line doesn't grow well suddenly, May i know the possible reasons for that?

How to compare pre-trained models with traditional models in the field of NLP?

Is it possible for two human neuronal cell lines to have different gene sequence for the same protein?

Genome-wide gene identification protocol ?

Any specific pipeline or parameter to analyse total transcriptomics study of brain tissue samples?

RNA-seq DEGs downstream analysis and reporting in literature?

Could you please suggests a tutorial or paper or any other materials to learn multiple gene expression in TCGA and GTEx databases?

Are there any databases for gene co-expression or expression pattern clustering?

Can I arbitrarily take a cycle number of 40 for the gene that was not expressed at all in qPCR results calculation?

Which CpGs have a significant effect on gene expression?

Found that there was a multiple band when amplifiying targeted gene ?

Why is there no protein expression but antibiotic resistance in transduced A549 cells with pLVX-TetOne-Puro lentivirus?