I am reporting the results sections of different papers and I need to know the unit of measurement of PCR. Is it enough to report if the expression increased or decreased (binary)? If so, can I mention in binary outcome how much fold it increased or binary should not have numbers? Most of authors just mention increase/decrease expression and very few mention by what percent or by how many folds. Thank you!
Hi Heba
it depends on your application. Some PCRs are done just to confirm the presence or not of a product, i.e. has it worked or hasnt it; Binary in your words
For expression data however, a quantitative rather than a simple qualitative answer always looks better and is more informative. In the past this was based on densitometry of agarose gels but correcting or normalising with a housekeeping gene: That is expressing fold change by citing the ratio of band density for your gene of interest compared to your housekeeper. However, in terms of dynamic range gel saturation etc this has limitations so today if people are interested in expression (quantitative) data this tends to be done by real time qPCR
In conclusion: For a figure in a paper or for something centrat to a story you are trying to tell expression data tends to be expressed quantitatively. For something preliminary or not central to a story you are trying to tell and not part of a figure in a paper or thesis etc then simply saying something has gone up or down (with a gel image to back that up; remember to show a normalised housekeeper however) i.e. a simple qualitative or binary answer is acceptable. It really does depend on context and what you are trying to show
Both techniques are semi-quantitative at best, this is why researchers usually refer to relative increases compared to a reference. Relative measurements are always easier than absolute measurements, even with highly quantitative enzyme assays or ELISA tests. Even in industry, absolute values are not very common, for instance if you want to buy the enzyme alpha amylase from different companies, you cannot compare them by trusting their units and dividing by the price to get the best deal, you have to buy a sample from each company, test them in your particular application, and divide your result by the price to find the best deal.
RT-PCR is particularly unreliable (compared to a good old Northern Blot) because you always get sharp bands and you cannot check if your RNA is slightly degraded, mostly degraded or undegraded. Both techniques are based on stepwise amplification of a signal, so that increases the noise. A reproducible factor 2 increase or decrease is hard to verify and I consider this the limit of resolution, smaller changes (i.e. factor 1.5) may be published but I believe none of it.
Best wishes,,
Jurgen
Thank you so much! It is not paper or thesis so I will use binary measures. Can I report that expression increased by 20% for example? Or in case of binary outcome, I don't mention the percent of increase?
If expression increases by 20%...that's only a factor 1.2. A two-fold increase is a 100% increase, and that's what I consider the minimum change to be trusted. When you do biological replicas (extract new protein from new samples, or new RNA from new samples), the variation is usually higher than a factor 2 and you need lots of repetitions to validade your factor 2 increase (100% increase). Doing a technical replica (repeating the PCR or the western on the same extract) is not really a repetition, and even then you can get quite variable results.
- Badges
- Science method