The dCt value is something like an expression value, normalized to some internal reference gene. A dCt value itself has no particular meaning, it is just some arbitrary value.Therefore, a dCt value will not tell you anything about a regulation of a gene (it even does not tell you much about its expression in any meaningfully interpretable way either, because it is relative to some arbitrary measure of some arbitrary reference gene).

BUT: dCt values from different samples/conditions can be compared. This is the key point. Differences in dCt values indicate differences in the expression (or your gene of interest) between different samples/conditions.

So whatever regulation or differential expression you are interested in, is MUST eventually determined from the comparison of different dCt values, i.e. by the difference between dCt values - and such differences between dCts values are ddCt values. In contrast to the dCt value, ddCt values do have a meaningful interpretation: a ddCt is the log fold difference (or log fold.change) in expression between the compared conditions/samples.

To analyze the differential expression (regulation), there is no way around using ddCt values. Most importantly, any uncertainty about the differential expression must be evaluated using the ddCt values. This means the relevant statistics to judge your results are the ddCt value and its standard error or (better!) confidence interval. Many people show standard errors (or confidence intervals) of the dCt values, but they do NOT contain the relevant information about the uncertainty in the estimation of the ddCt (the differential expression).

But a comment on Mikhails answer:

Since the dCt value itself is arbitrary, showing (or reporting) a dCt value is not very instructive by itself. BUT there is an exception, where showing dCt values really makes sense: when you have several conditions and when you want to explore the differential expression beween all pair-wise combinations of these conditions. It can also be useful when you have some time series or a concentration series and you want to analyze the change in expression over time or depending on the concentration. Nevertheless could here also ddCt values be given directly. I would prefer seeing ddCt values (with a measure of their uncertainty!) because this is the eventually relevant information, and it is better to give this information explicitely instead of forcing the reader to eyeball it from figures or numbers showing the dCt values.

And a second comment:

A dCt value by itself does not indicate a high or low expression, because its value depends on the reference you selected. To judge whether the expression of your gene of interest is rather high or rather low you will need to look at the Ct values of this gene directly, and you need to take into account what amount of biological material (roughly: the cDNA of the RNA eqivalent of how many cells) you used in your PCR. The result can only be approximate - you can distinguish "rather high" and "rather low" expressions (but there is a very large dynamic range, though!). It can't be a precise determination unless you use external standards and possibly spike-in controls to control for sample preparation and RT efficiency and all that.

Hi Wiyada! I suggest you to use the 2^ddCt method,  since this formula allows you to report gene expression of your group of interest (stress, in your case), relative to a calibrator group (normal condition). According to the formula, the "normal group or calibrator" will allways be normalized to a value of 1, because 2^(0) = 1.  You will have a direct comparison of how much is increased or decreased your gene of interest in comparison to the normal condition.

You can also use the 2^dCt method, and overall you should obtain a similar pattern of expression if there is a significant difference, but according to your experiment, i think the 2^ddCt is more appropriate.

How this help you, good luck !!

I personally prefer  2^dCt formula because it gives more information than 2^ddCt. First, it shows gene expression level before and after stimulation, for example 10 copies before stimulation and 50 copies after stimulation. You may easily calculate a 5 fold induction. In the case of ddCt method, you will see only 5 fold induction without knowing the levels of expression. In our experiments we isolate human monocytes from blood and find them sometimes activated and not responded well to stimuli. dCt but not ddCt method can indicate for activated monocytes. 

So, it is up to you what method to use. 

The dCt value is something like an expression value, normalized to some internal reference gene. A dCt value itself has no particular meaning, it is just some arbitrary value.Therefore, a dCt value will not tell you anything about a regulation of a gene (it even does not tell you much about its expression in any meaningfully interpretable way either, because it is relative to some arbitrary measure of some arbitrary reference gene).

BUT: dCt values from different samples/conditions can be compared. This is the key point. Differences in dCt values indicate differences in the expression (or your gene of interest) between different samples/conditions.

So whatever regulation or differential expression you are interested in, is MUST eventually determined from the comparison of different dCt values, i.e. by the difference between dCt values - and such differences between dCts values are ddCt values. In contrast to the dCt value, ddCt values do have a meaningful interpretation: a ddCt is the log fold difference (or log fold.change) in expression between the compared conditions/samples.

To analyze the differential expression (regulation), there is no way around using ddCt values. Most importantly, any uncertainty about the differential expression must be evaluated using the ddCt values. This means the relevant statistics to judge your results are the ddCt value and its standard error or (better!) confidence interval. Many people show standard errors (or confidence intervals) of the dCt values, but they do NOT contain the relevant information about the uncertainty in the estimation of the ddCt (the differential expression).

But a comment on Mikhails answer:

Since the dCt value itself is arbitrary, showing (or reporting) a dCt value is not very instructive by itself. BUT there is an exception, where showing dCt values really makes sense: when you have several conditions and when you want to explore the differential expression beween all pair-wise combinations of these conditions. It can also be useful when you have some time series or a concentration series and you want to analyze the change in expression over time or depending on the concentration. Nevertheless could here also ddCt values be given directly. I would prefer seeing ddCt values (with a measure of their uncertainty!) because this is the eventually relevant information, and it is better to give this information explicitely instead of forcing the reader to eyeball it from figures or numbers showing the dCt values.

And a second comment:

A dCt value by itself does not indicate a high or low expression, because its value depends on the reference you selected. To judge whether the expression of your gene of interest is rather high or rather low you will need to look at the Ct values of this gene directly, and you need to take into account what amount of biological material (roughly: the cDNA of the RNA eqivalent of how many cells) you used in your PCR. The result can only be approximate - you can distinguish "rather high" and "rather low" expressions (but there is a very large dynamic range, though!). It can't be a precise determination unless you use external standards and possibly spike-in controls to control for sample preparation and RT efficiency and all that.

Thank you all,

Zyanya Reyes-Castillo;

Mikhail A Gavrilin;

Jochen Wilhelm, 

Thank you for the useful comments,

Now, I have many samples such as wild-type, mutant and complemented strain. Hence, I would like to compare the interested gene expression under normal and stress condition between wild-type, mutant and complemented strain. 

Is it better if I use 2^dCt for comparing the gene expression?

ฺBest Regard,

Wiyada Dankai

Again:

A dCt value itself is related to the expression. To compare expressions, you comapre dCt values. Such a "comparison" is mathematically the difference of two dCt values, what is a ddCt value. A ddCt value is a log fold-change. You can further compare ddCt values (for stress-induced changes in gene expression) between different strains. This comparison is mahematically again a difference, now between two ddCt values (one may call this dddCt values).

Given you calculate dCt = Ct[ref] - Ct[goi],

you can compare dCt[stress] to a dCt[normal], and if dCt[stress] > dCt[normal], the gene of interest was induced by stress, if dCt[stress] < dCt[normal] then the gene was repressed by stress. That's as simple as that. The difference between these values: ddCt = dCt[stress] - dCt[normal], gives you the directtion and the amount of the stress-induced gene regulation. The sign gives you the direction (positive: induction; negative: repression), and the (absolute) value gives you the strength or amount of regulation (the "effect size" of the factor "stress").

Thank you, Dr.Jochen Wilhelm

I would like to ask you following; 

For example; 

The Ct of Ref wild type (normal condition) is 12 

The Ct of Goi wild type (normal condition) is 21

The Ct of Ref wild type (stress condition) is 14

The Ct of Goi wild type (stress condition) is 20

Whereas, 

The Ct of Ref mutant (normal condition) is 21

The Ct of Goi mutant (normal condition) is 30

The Ct of Ref mutant (stress condition) is 16

The Ct of Goi mutant (stress condition) is 23

Can I use the formula  dCt = Ct[ref] - Ct[goi], and then use the 2^-dct ?

Best Regard,

Wiyada Dankai

Jochen,  I agree that ddCt gives you direction of the gene regulation. What I want to point out that dCt gives you start and end points in this gene regulation and you always may see this direction, calculate it and have more control on your experiment. Here is an example from our work: we isolated monocytes from patients during several days and stimulated them with LPS. First we used ddCt method and observed an increase in ddCt value for IL1B (meaning induction) and then, at one time point, there was no any induction. How to explain this? We decided to use dCt method - we use RCN ( relative Copy Number). RCN = (E^-dCt)x100. E=2, we normalize to an average expression of two housekeeping genes and our gene expression (or relative copy) was related to 100 relative copies of average 2 housekeeping genes. We were able to separate values before and after stimulation and what we have found: Not activated monocytes show expression for IL1B around 1-5 RCN and 100-200 RCN after stimulation with LPS. However, unstimulated samples from day X showed RCN around 200 and addition of LPS didn't change this gene expression. It suggested that monocytes were already highly activated and did not respond to the LPS stimulus. Then we found that this patient had secondary hospital infection and the same day X. Unsung RCN, based on dCt we were able to see activated monocytes and explain our observations  and this would not be possible with ddCt method. We use this method for more that 10 years, describe it and publish in all our publications. We know RCN values for more that 100 genes were are working with and may better control our experiments and status of human samples.

Of course, absolute quantification is the gold standard. It gives you starting and end points, what our relative quantification is mimicking. 

Dear Wiyada,

It is important to note that  "2" is only a theoretical qPCR efficiency; it is assumed that you will have doubled your DNA each cycle. Though a good approximation, it is rarely the case. Typical (good) efficiencies range from 1.8-2. Assuming that you will have an efficiency equals to 2 (and the same efficiency through all samples or conditions) introduces error. To better calculate efficiency, you can analyse your data using the free software LinReg. It is easy to use and after spending a little time learning how to use it, you will achieve more accurate results and more realistic values of efficiency.

You can also use a specific free software to do the statistics from your qPCR experiments: fgStatistics, search for it in the internet. It returns you the p-values for each comparison.

Also, I would suggest you that in order to find a suitable housekeeping gene for your conditions/tissues/treatments, try 3-4 candidates and then find the more stable one by using Bestkeeper, GeNorm, Normfinder or other available software.

It is very important too to have several biological replicates (often, 5-6 replicates are sufficient) for each treatment/control in order to obtain robust and more reproducible results; and to obtain a good p-value too.

@Wiyada:

Wildtype, normal: dCt = 12-21 = -9

Wildtype, stress: dCt = 14-20 = -6

Wiltype stress-sffect: ddCt = (-6) - (-9) = +3 (induction, about 8-fold)

Mutant, normal: dCt = 21-30 = -9

Mutant, stress: dCt = 16-23 = -7

Mutant stress-effect: ddCt = (-7) - (-9) = +2 (induction, about 4-fold)

Interation of Genotype x Stress: ddCt[Mutant] - ddCt[Wildtype-] = 2 - 3 = -1

--> stress-effect in mutants is about half of the effect in wild-types.

What is improtant is some indication of uncertainty (what is completely missing in this example)!

.

@Mikhail:

That's a good point and a good example why  and when it is worth to look at the expression and not only at the regulation. I don't see any advantage of using RCN values instead of dCt values, though. Both are "arbitrary" measures and you need some experience to interpret these values in the context of your assay/system. The addition calculations required to get the RCN values seem superfluous to me.

Thank you all, 

There are many useful comments.

Thank you, Dr.Jochen Wilhelm 

Thank you that you examine how to calculate.

I hope the result like this (the interested gene lower express ever stress) because my idea geneA in wild-type can regulate the interested gene expression. So, the mutant has a defect of geneA that why the mutant should low expression.

Jochen,

all we need is to put the data to excel spreadsheet. We usually calculate dCt, ddCt and then RCN. So we have all desired information. Hovered, I prefer (my choice) to express the data as RCN (yes, arbitrary, relative to expression of housekeeping genes) because it not only tells me the direction and fold of gene regulation, but also starting and end values (yes, relative). All the values are very reproducible. If you (and Wiyada as well) want, I may send my excel spreadsheet by mail.

Thank you Dr. Mikhail A Gavrilin

Please sent me for [email protected] or [email protected]

I think many comments can improve my work.

Best regard

Wiyada Dankai

Dear Mikhail A Gavrilin,

Could you please sent to me the excel sheet as well.

Many thanks,

Dalia Sabrei

Sorry, I forgot to post my email, [email protected] 

Many thanks,

Dalia Sabrei

Done

Hi Mikhali, I would also verzy much like to use your spreadsheet. Would you share it once more?

Many thanks and best from Boston,

Hannes

Dear Mikhail A Gavrilin,

Could you please send to me the excel sheet as well.

Thank you very much e-mail: [email protected]

Hannes and Jorge, I will do it on Monday. Feel free to ask any question about it.

OK, Thank you very much..  Best regards

Dear Mikhael, sorry to ask again but could I please have that spreadsheet too?

my e-mail is [email protected]. Many thanks in advance

Sent

Dear Mikhail,

I have stumbled across this conversation from over a year ago. Would it be possible to also have this spreadsheet? Thanks!

[email protected]

sent

Dear Mikhail

I m trying to quantitate gene expression in stem cells without having any controls to do relative gene expression.Its sounds interesting, how you managed to performed IL1B expression in monocytes. Will you please send me the excel sheet. I have similar condition like you mentioned. I ill try to use your way of analysis.

Thank you for sharing.

Sorry i forgot to add my email id : [email protected].

Thanks

Sent. Excel template and Power Point presentation.

Hello Mikhail,

If you see this, would you please send your spreadsheet along to [email protected].

Thank you!

Ok. Tomorrow

I prefer 2^dCt to study the gene expression, it provides the actual level of gene expression. The key point is using stable reference gene to have data robust.

2^dCt is a measure relative to a loading control / reference gene. It does not represent the "actual expression level". Knowing that 2^Ct is 152, say, doesn't tell me anything. It does neither indicate a high nor a low expression. This value, taken in isolation, has no particular meaning (unless I would know the expression level of the reference gene, what I usually don't know!).I can only compare two such values from different reactions/samples to see in which sample the expression was higher (or lower) - and this information is also provided already by the dCt values, except that dCt values can be compared by their difference whereas 2^dCt values have to be compared by their ratio.

You don't get any additional information about the expression level just by transforming a number (e.g. raising it to the power of 2). Where should that additional information come from?

Dear Dr. Mikhail A Gavrilin,

would you share your excel sheet with me as wel?

e-mail: [email protected]

Thanks a lot!

Dear Dr. Mikhail A Gavrilin,

would you share your excel sheet with me as well?

e-mail: [email protected]

Done

Mikhail A Gavrilin

Dear Mikhail A Gavrilin, COuld you please send me the excel sheet template? my ID is [email protected]

Dear Mikhail A Gavrilin ,could you please share your excel sheet with me. kindly i need your consideration. Thanks

my email is [email protected]

Basirudeen and Shamima,

i am on vacations and will be back to work on August 5th. I will try to remember to send it to you but please feel free to remind me.

regards

Dear Dr. Mikhail A Gavrilin,

would you share your excel sheet with me as well?

e-mail: [email protected]

Sent

Dear Dr. Mikhail,

Could you please share your excel sheet with me as well?

e-mail: [email protected]

Dear Mikhail A Gavrilin

Could you please share your excel file with me ?

[email protected]

Thanks

Dear Mikhail A Gavrilin

would you share your excel sheet with me as well?

e-mail: [email protected]

Thanks

Sent

Dear Mikhail A Gavrilin

Will you also share your excel sheet with me?

e-mail: [email protected]

Done