I would like to add a point: there are markers that are not just absent or present on a cell - they can be expressed in different numbers. WIth these kind of markers we generally define a negative population, a low or dim population and a high or bright population. The low or dim populations can have different properties as compared to the negative or bright ones. You would be able to tell the difference by looking at densitiy blots.

Basel, your question is very well taken. Positive/negative is correct, as negative/dim/bright expression. Alternatively, medium and high can be accepted too. I do not approve "low", because it implies that the antigen is expressed at low levels, but expressed. "Low" is mostly used by people doing mouse immunology and means "negative" (unless it means "too low to be detected"). However, as you pointed out, it's negative by FACS. So, "negative" is more appropriate, if you refer to a population not expressing the antigen.

Dim/bright or medium/high are relative and depend on where you place your gates (or if you have two different, positive populations expressing different amounts of the antigen).

Thank you very much, Enrico. Your answer is pretty clear and informative.

One of the problems with flow cytometry separations is that they depend on how the gates are set for each color and this can vary from individual to individual and between labs using the same model machine, leave alone different brand machines. For DNA content and DNA synthesis marking/separation, the problem is not very large since one can set the gates on a 2n DNA content and a 4n DNA content and then separately detect the DNA synthesising cells with BrdU antibody. But for cell surface markers, the problem of how to get the positive and/or negative population(s) with high purity is far less clear. This makes the results somewhat qualitative even though one can cite a specific number or fraction of cells selected by one's gating parameters. An alternative method is to use the same antibody selections but bound to magnetic beads and then doing magnetic bead based separations. The advantage of this method is that the magnets make the decisions for you.

Christopher, thank you for your answer. Though, I am from the computational side of computational biology. So, I do read loads of pure biological papers related to my specific research to biologically analyse my computational results, but I don't really design or implement any of such biological experiments. My main issue was simply about "reading" these plots which had something ambiguous for me. Anyway, your answer has surely provided me with useful information about these techniques.

I would like to add a point: there are markers that are not just absent or present on a cell - they can be expressed in different numbers. WIth these kind of markers we generally define a negative population, a low or dim population and a high or bright population. The low or dim populations can have different properties as compared to the negative or bright ones. You would be able to tell the difference by looking at densitiy blots.

I would add that usage of hi/lo/mid designations is simply done in order to stray away from absolutes. As some have pointed out, physiological differences between dim/mid and hi/+ populations can be resolved. Furthermore, unlike very specific antigens like CD4 and CD8, many surface antigens cannot be resolved into quadrant gates by 2-dimensional plots, requiring analytical descriptions such as lo,hi,mid, etc. Nowadays, using "logicle" (or bi-exponential) gating gives rise to defining populations below background, requiring sometimes the use of descriptors that are not purely binary.

for a good review on how gating changes things, see "Interpreting flow cytometry data: a guide for the perplexed" Nature Immunology 2006

This discussion helped me, thank you all - Reinaldo

How we must determine high expression of a surface marker in flow cytometry?