@Felix : you are right, this is why I proposed to use "RNA abundance" or "transcript level" rather than gene expression.

@others: RT-PCR does not measure transcription but a sum of processes including transcription.

I think that gene transcription would be even less appropriate.

The level of mRNA is the result of complex processes, including transcription, but also mRNA stability / decay, post-transcriptional modifications etc...

I would stick to gene expression, or if you want to be more accurate, level of transcript or RNA abundance.

@Vincent. Thanks for your view, but could you substantiate why we should stick to gene expression since by using RT-PCR we have not gone all the way to see the protein, which is, as I understand gene expression.

By avoiding the term "gene transcription" we avoid the complexities you mentioned, on the hand by using gene expression we are over-stating the RT-PCR results because there is no protein involved.

Technically gene transcription is the more descript/accurate term, but it is almost universally referred to as gene expression.

@Guy. It is this universalism here that I am at loggerheads with. Even if it is universally used but inaccurate, should it be continued? I think in science we cannot afford to be inaccurate.

@Felix : you are right, this is why I proposed to use "RNA abundance" or "transcript level" rather than gene expression.

@others: RT-PCR does not measure transcription but a sum of processes including transcription.

Vincent is absolutely correct. RT-PCR only measures RNA levels. However people often use changes in levels to infer changes in gene expression. You may know the biology of your gene sufficiently to determine whether that inference is justified or not.

@Michael. Tobias, an earlier contributor to the discussion shows the widespread use of "gene expression and RT-PCR". I believe in all those uses the authors do not work on inference but state it is gene expression. If we worked on inference we would have to provide the justification to convince everyone that had we looked for protein we would have found it.

Therefore, should there be definitions for the bench scientist (which do not have to be accurate because they will assume) and definitions for learners e.g., students (where accuracy of definitions is of paramount importance)?

Gene expression is a process comprising many steps. Yet, a measurement on any level of this process is a measurement of gene expression. Therefore, RT PCR does measure gene expression and most importantly changes in RNA levels can be called changes in gene expression.

However, and as many already pointed out, RT PCR does NOT measure transcription if applied to a steady state situation as performed in most of the cases.

In my opinion for an RTPCR assay is more commonly used , analysis of gene expression at RNA level. In the case of translation, for example in a Western blot assay, gene expression at protein level.

Many years back a reviewer criticized my paper for using the term "Protein Expression". I learned that "protein expression" is bad molecular biology slang. For example in Molecular Biology of the Gene (Watson), you will never once see the term "protein expression". Proteins are not expressed. Proteins are translated or synthesized from RNA of genes that have been expressed or transcribed. Thus RT-PCR is measuring gene expression. The "protein expression" slang now dominates the biotech world and it is an exhausting and pointless exercise to correct others. However, one can always write correct terminology in their own works and get satisfaction from that.

@Jeffrey. So I take it you are really not satisfied with the "protein expression" in the name of your Lab :) Yes it is about correct terminology.

My whole point of bringing this up was, students on an assignment peruse scientific papers and find all this molecular biology jargon and are confused as to what is the correct terminology. And explanation is not going to be easy, as you can read from the discussion here that not all agree as to what is the meaning of RT-PCR results. Are they gene expression or gene transcription? One can dismiss your gene expression meaning in RT-PCR on the same grounds others had dismissed protein expression. They all seem to be generalizing (gene expression) or abbreviating (protein expression) the definitions.

It is true sometimes gene expression and gene transcription are interchangeable and sometimes I use both only not to repeat the same words in the next sentence. However, I think there are some differences for the reasons raised above. If I evaluate mRNA level by RT-PCR, northern, RNase protection assay, microarray, etc…or any method that measures RNA steady state levels, I will be talking about gene expression because at the time of my quantification the gene may have been shut down-but if I use run-on, RNA pol ChIP, etc…for example I can talk about gene transcription because it is happening at the time of my quantification. This is very important to bear in mind because some regions of the chromatin are continuously transcribed but they are not expressed because the endogenous RNAi pathway degrades them into siRNAs. So, if you check by normal RT-PCR you may conclude that these regions are not expressed. Yet, they are transcribed. So, I think it comes to the context and what one is concluding from that.

I think protein expression is also a correct word. I will understand different things if you tell me that a cell doesn't express a protein versus a cell which doesn't translate/synthesize a protein.

@Abdelhalim. I think you contradict yourself in your answer. You say ... some regions of the chromatin are continuously transcribed but they are not expressed... And yet you say in your writting your interchangeably use these terms, which according to you are different process that occur independently and so are not one and the same thing.

I have to chime in support for the statement that RT-PCR only measures RNA levels (in case one has insured there is no DNA template in the mix…) Generally transcription correlates directly with gene expression (I don’t recall reading about cases when there is inverse correlation), but there has to be a very good understanding what exactly is the exact relation (value) for the gene of interest. Unfortunately it may not be the same for genes even in the same operon. So, "level of transcription" should be far more accurate.

@Felix You can’t take me on every single word. Yes, sometimes they are used interchangeably but within a context. Let just put the example of exogenous siRNA against a target gene in mammalian cells. I think I can say I block the expression of the gene but not I block the transcription of the gene. PTGS versus TGS. Literally for me, gene transcription refers to the productive and non- productive elongation processes by any RNA polymerase while gene expression refers to its integral and functionally relevant product only. From this, I can say partially transcribed gene but not partially expressed gene because people (at least for me) will understand two different things. Of course, I may be mistaken on this.

Definitely, you can not measure gene transcription by RT-PCR.

Transcriptomics (referring mainly to cytosolic RNA 'profile'), genomics (referring to nuclear DNA content), and proteomics (referring to protein 'profile') may be terms that could help out here. So, when performing qPCR on cDNA and/or RT-qPCR (in one-step fashion on RNA), one is measuring the "transcriptomic levels" of RNA/cDNA. Semantics and misunderstood, closely-related abbreviations confuses things, agreed. We all need to put on the same virtual headset of universal language, someday. Genome, transcriptome, proteome...

Increasing the level of granularity in the terminology, we might arrive at something like: genome, SNPeome, CopyNumberVariome, mirTronome, epigenome/histeome, topochromeome, nucleome, spliceome, exome, miRnome, transcriptome, proteome, golgi/post-translateome, metabolome, secretome ... and so on(eome)... and the obligate associated international data bases will responsibly follow. In addition, assuming that Moore's law holds true (regarding the promised exponential growth of computing power over time); the evolution of such data bases should be simple-simon. It will all be accessible information.

How expansively we name the system(s) we study directly affects the nature and scope of the very questions we ask in the first place. If our view(s) of the systems we study are overly-simplistic, we run the risk of asking questions that are similarly myopic (although ignorance and bliss have long afforded us magnificent scientific discoveries, regardless).

But, if we acknowledge the intrinsic (actual) extreme complexities (known and unknown) up front, the more informed our initial questions can be.

What level of granularity are you/we//they comfortable with? How much sight is too much, and how much blindness is not enough?

If 'omics' is/are the new discipline(s), then improving algorithms on all related fronts consititutes the palpable infantry of "panbiomics" these days. One recent shining example of this would be:

A new unsupervised gene clustering algorithm based on the integration of biological knowledge into expression data

BMC Bioinformatics 2013, 14:42 doi:10.1186/1471-2105-14-42

Marie Verbanck ([email protected])

Sébastien Lê ([email protected])

Jérôme Pagès ([email protected])

Attempts at putting the 'entire picture' together (from the genome/epigenome to the 'secretome') will invariably be championed by this and similar efforts on into the near and distant future. Open eyes, open hearts, open minds, or eyes wide shut. Perhaps it is high time to jump aboard, everyone; things only get more complex from here, language barrier or not...

For the sake of future demonstration:

Imagine the time when the term "epiproteome" will finally belong itself into everyday vernacular for: "proteins that are expressed in progeny in remarkably near epi-Lamarckian-fashion." Such things are sure to make the cover of Newsweek, and perhaps will go on to spawn an iterating escapde of derivative pop-culture novels (as has been human custom). Science fiction has often preceded things that are 'fully' understood and incorporated as common-knowledge later (much like the double-strandedness of DNA being common knowledge now and not before 1952). Myth begets science continually (always has) -- 'tis the human way about things?

And yes: we should be very careful (and specific) in the words we use in all languages to define/describe exactly what we are talking about. Music demonstrates in 12 basic tones that this is possible, yet language still struggles as if in a lonely wheelchair over there, somewhere, in a dimly-lit corner of the human room too few tend to visit.

Terminology for qPCR/RT-qPCR has been addressed by the MIQE Guidelines since 2009. It is an on-going effort that depends upon the speed of the internet, the resistance of the human skull, and upon interpersonal trust as to how much time it will take for universal acceptance. Some good ideas never catch on.

One example: "Ct" threshold cycle, has now been re-named "Cq" for "quantification cycle." It is a better term, and is more definite about what it actually stands for in all languages, yet, "Ct" still persists because it was the among the first names people fell in love with, and because it was somewhat intuitive.

One day, perhaps our heiroglyphs will all converge into universal tongue. I couldn't agree with the initial question posted by Felix Toka here more. He is frustrated by the English language in these matters, and so am I.

Additionally, the over-use of 3, 4 or 5-lettered acronyms these days is public enemy #1 in the house of understandable scientific discourse. Many get lost and tune out due to such brevity. Creation of clearer definitions of terms and clearer use of abbreviations as time proceeds, seems a highly worthy idea.

Summarizing briefly: gene expression covers both transcription and translation processes. With RT-PCR you measures mRNA results, which can be affected with either of them.

qPCR/RT-qPCR provides one part of the 'gene expression' picture, proteomics provides another, miRNA provides another, there is also microDNA involved as well. The term "gene expression" is evolving these days - it is beginning to include more things, that, before 15 years ago were as yet unknown and or poorly understood. The final "language" of DNA is protein, so, in a semantic sense, the ultimate goal of DNA is to 'express' itself in terms of protein...

Hi Jack

Very nice essay. You think science is driven by Lamarckian forces or am I mistaken? I need to think about it. I suggest you join a thread at https://www.researchgate.net/post/Is_there_a_scientific_alternative_to_neo-Darwinism_for_understanding_biological_evolution

The perception that the word gene covers only proteins-coding genes is not accurate. There are a lot of non-coding RNA genes, such as rRNAs, tRNAs,…..etc (ncRNAs) whose transcription/expression is also covered by the terminology.

Also, Jack, The DNA can “transcribe” itself but sometimes it can’t “express” itself when something interfere in between …like a censorship (you talk but nobody can hear you if you voice as been distorted, etc…)

Apparently I have confused you - I don't ascribe to Lamarckian theory. When methylation patterns are carried over into progeny (not being totally wiped clean from parent to child), that mimics a pseudo-Lamarckian idea but not the Lamarckian debacle proper. Just pointing out the strange near-parallel is all. I expect that all posts may get misconstrued depending upon the reader, don't you? Thanks for your thoughts.

To shortly answer your primary question without going too much into details: you will surely make a correct statement if you use “ XXXX mRNA levels” (where XXXX is the name of the gene/protein from which you are interested).

I am surprised by how far the discussion went, it turned out to be very interesting, but definitely qRT-PCR does not represent transcription levels. If you want to mesure transcriptional activities , you will have to go through different apporaches such as Run On assays. qRT-PCR analyses mRNA levels at T time, which will be the results of two processes: synthesis and degradation^^

And I know of genes where you could have very intensive transcription with a very low amount of RNA....so to make it short, qRT-PCR is in no way gene transcription but just mRNA levels.

Transcriptome: The complete set of RNA transcripts produced by the genome at any one time. The transcriptome is dynamic and changes under different circumstances due to different patterns of gene expression. The study of the transcriptome is termed transcriptomics. RT-qPCR may therefore be said to probe the transcriptome.

Transcript level is essentially correct. Depending on what you use as a reference you might call it transcript ratio or concentration. Ideally, you could measure absolute numbers of your template in an appropriately standardized assay.

Transcript is better than mRNA, because you may interrogate several species including non-fully processed ones, depending on your primers.

You do not measure transcription rates by RT-PCR (unless you set up specific assays for that purpose). Rather, the transcript level is a function of transcription rate and turnover (as well as splicing in some cases). Depending in particular on the turnover rate (!) the amount of mRNA at any specific time may have little to do with the rate of transcription. So gene transcription is almost always the wrong designation.

Gene expression actually has a much broader meaning than mRNA level, but if the context is clear in which you use the term, it should be fine as a generic term.

So that is what I keep telling my students. Please, everybody correct me if I am wrong...

WAS

There's no need to get into the "-omics" here - RT-PCR isn't measuring "transcriptome levels", it's measuring transcript levels.

RT-qPCR (reverse transcription quantitative PCR) therefore may be said to probe the transcriptome for transcript levels at singular time T, as an indication of gene expression at the mRNA level. This however assumes that pre-mRNA isn't being co-isolated and co-investigated (inadvertently) as well.

Wolfgang Schulz has said it well: "Gene expression actually has a much broader meaning than mRNA level, but if the context is clear in which you use the term, it should be fine as a generic term."

Transcription level of pre-mRNA and mRNA - or just cytosolic mRNA? The term/phrase "'transcript level" thus has this limitation, but, I think it is the best we can do without necessitating the addition of more words and/or descriptors to help identify [more exactly] which "gene expression" 'compartment' we are referring to...

"Level of gene transcripts" ambiguously refers to pre-mRNA and mRNA by default does it not? So, it would always seem to be important to use primers (or primers and probe) during RT-qPCR that always span an exon-exon junction when probing for transcript levels in the 'cytosolic compartment.'

If this is not done, then nuclear pre-mRNA is always a part of the RT-qPCR signal, making it less truly indicative of the transcripts potentially on their way to be translated/expressed as protein.

In eucaryotes (eukaryotes), 5-10% of total RNA is thought/estimated to be nuclear RNA, and some of that nuclear RNA is pre-mRNA. Pre-mRNA transcripts (prior to splicing) run the risk of being co-amplfied (as does contaminating gDNA) when permissive primers/probes are used for RT-qPCR. The term "transcript level" should always be further clarified/described/qualified based on the specifics of the RT-qPCR method(s) actually used.

Further, ample DNase treatment of total RNA samples subjected to RT-qPCR is part of the key to solving the gDNA contribution to spurious signal, but, it doesn't solve the problem of pre-mRNA, which is still there...

If ~2 to 2.5% of total RNA is mRNA, and 5-10% of total RNA is nuclear RNA, then if pre-mRNA is a credible/influential portion of total nuclear RNA (on RT-qPCR analyses), think of the dilema that all RT-qPCR has always been faced with...

Unless primer-probe designs always span exon-exon junctions (if they are available), RT-qPCR analyses can always be slightly or largely misleading depending on rate of turnover or degradation of pre-transcripts at the nuclear, pre-mRNA level.

Compartment, compartment, compartment...

Procaryotes (prokaryotes) would fare worse in this argument.... having a paucity of difference between nuclear and cytosolic compartments... sequence-wise.

There are many answers. I think that Jack Gallup'answer is the best abstract because his compilation is very instructive.

Transcript levels. Gene expression is vague and while we generally infer that this means gene--> protein, then what do we do with RNAs that do not make a protein (i.e. H19)? Keep it simple. Transcript levels, protein levels, etc.

So, to be clear: pre-mRNA + mRNA (translated or not) = transcript level.

But, I still think the original question begs further clarification in the answer: and that is whether or not the RT-qPCR (reverse transcriptase quantitative PCR using RNA in a two-enzyme [RT and 'Taq'] mastermix) is carried out responsibly using primers that span a genomic intron-exon (mRNA exon-exon) junction. In a general sense though, I do agree with David Sassoon, and perhaps that is all the original question really needed for an answer.

There are additional concerns however, as I've attempted to point out.

@Vincent and Michael: "RNA levels" and "RNA abundance" are still are not truly descriptive enough...

So I repeat my prior post here:

Transcription level of pre-mRNA and mRNA - or just cytosolic mRNA? The term/phrase "'transcript level" thus has this limitation, but, I think it is the best we can do without necessitating the addition of more words and/or descriptors to help identify [more exactly] which "gene expression" 'compartment' we are referring to...

"Level of gene transcripts" ambiguously refers to pre-mRNA and mRNA by default does it not? So, it would always seem to be important to use primers (or primers and probe) during RT-qPCR that always span an exon-exon junction when probing for transcript levels in the 'cytosolic compartment.'

If this is not done, then nuclear pre-mRNA is always a part of the RT-qPCR signal, making it less truly indicative of the transcripts potentially on their way to be translated/expressed as protein.

In eucaryotes (eukaryotes), 5-10% of total RNA is thought/estimated to be nuclear RNA, and some of that nuclear RNA is pre-mRNA. Pre-mRNA transcripts (prior to splicing) run the risk of being co-amplfied (as does contaminating gDNA) when permissive primers/probes are used for RT-qPCR. The term "transcript level" should always be further clarified/described/qualified based on the specifics of the RT-qPCR method(s) actually used.

Further, ample DNase treatment of total RNA samples subjected to RT-qPCR is part of the key to solving the gDNA contribution to spurious signal, but, it doesn't solve the problem of pre-mRNA, which is still there...

If ~2 to 2.5% of total RNA is mRNA, and 5-10% of total RNA is nuclear RNA, then if pre-mRNA is a credible/influential portion of total nuclear RNA (on RT-qPCR analyses), think of the dilema that all RT-qPCR has always been faced with...

Unless primer-probe designs always span exon-exon junctions (if they are available), RT-qPCR analyses can always be slightly or largely misleading depending on rate of turnover or degradation of pre-transcripts at the nuclear, pre-mRNA level.

Compartment, compartment, compartment...

Procaryotes (prokaryotes) would fare worse in this argument.... having a paucity of difference between nuclear and cytosolic compartments... sequence-wise.

I believe these statements represent legitimate concerns. Although I do appreciate the brevity of your remarks - can you expand a little on what you are trying to point out - given the apparently popular recepetion of your comments?

Just because a response is popular, doesn't mean that it is right. It could be comfort level with the response to the topic at hand masquerading as truth - you know what I mean?

Please elaborate when you have the time.

I agree that transcript level is by far the better term, even if I must admit that I have used the term gene expression a number of times as well, particularly when I have used quantitative real-time RT-PCR to look the transcript level change in a series of time-points after treatment.

The most accurate description would acknowledge the limitations of the techniques employed instead of an all-inclusive term such as 'gene expression levels'. It could be something like 'polyA tailed, taq pol amplifyable' etc as that is what is being detected and not anything else.

I am confused now.. Helppp..

So, if I have carried out a simple RT-PCR using specific primers to detect my gene of interest (expression levels), would I write mRNA expression or gene expression (for a paper)?

Side note: I also carried out a ELISA, and with some of the genes of interest the results align with the RT-PCR data...

None,

As what you are looking at is the amount of RNA and not the transcription. You are looking at a steady state situation where the amount of RNA is depending of both its synthesis/processing and its degradation.

You will be surprised by the transcription level of some genes compared to the amount of RNA quantified.

For transcription it is a different kind of analysis. Therefore it is mRNA levels...transcript levels etc...

I know that usulaly we use the shortcut to 'gene expression levels'....but we are all off !

Thank you very much Pascale,

Kind Regards