If you are only using a single stain to look at your cells, then yes, unstained cells are your negative control. If you are adding more than one antibody (thus, more fluorochromes), you should add all your antibodies except your FITC-conjugated one. This is an FMO control (Fluorescence Minus One). Repeat this for every fluorochrome you use, to set your gates.

Do not ever use an isotype control to set your positive gates! (as suggested by Eiji Miyoshi) The only thing that an isotype control can tell you is if you have severe aspecific antibody binding (often Fc-mediated). You cannot use it to 'quantify' aspecific binding of your antibody, because your specific antibody and isotype control will have different aspecific binding properties by virtue of being different! Isotype controls are largely a thing of the past, and hopefully they will now stay in the past.

You can control background staining on negative cells by properly titrating your antibody.

Edit: I am wrong in saying that prof. Miyoshi suggests using an isotype control. He recommends an indirect staining protocol, which is something entirely different.

However, my comment about not using isotype controls as a gating control is still valid!

Yes, that is a good negative control for your experiment. I would use that negative control for each antibody you're using.

I fully agree with Svasti. You can use the population of unstained cells to set the threshold level so that the flow cytometer sends only events above this level to the computer. This helps you to reduce background by improving the signal/noise ratio.

When we use FITC-conjugated antibody in FACS analysis, we use unstained cells as negative control. But indirect staining with FITC-conjugated second antibody is recommended to avoid non-specific staining

Eiji

Thank you soooo much for everyone's kind advice. And I will try to do as above design. Thank you!!!

If you are only using a single stain to look at your cells, then yes, unstained cells are your negative control. If you are adding more than one antibody (thus, more fluorochromes), you should add all your antibodies except your FITC-conjugated one. This is an FMO control (Fluorescence Minus One). Repeat this for every fluorochrome you use, to set your gates.

Do not ever use an isotype control to set your positive gates! (as suggested by Eiji Miyoshi) The only thing that an isotype control can tell you is if you have severe aspecific antibody binding (often Fc-mediated). You cannot use it to 'quantify' aspecific binding of your antibody, because your specific antibody and isotype control will have different aspecific binding properties by virtue of being different! Isotype controls are largely a thing of the past, and hopefully they will now stay in the past.

You can control background staining on negative cells by properly titrating your antibody.

Edit: I am wrong in saying that prof. Miyoshi suggests using an isotype control. He recommends an indirect staining protocol, which is something entirely different.

However, my comment about not using isotype controls as a gating control is still valid!

Generally, your approach is OK. You will not be able to conclude that your antibody binding is specific, but often it is considered acceptable.

If you want to raise your level of perfectionism later (for instance, if you are using a home made antibody, not verified by supplier quality assurance, and you need additional data to prove it's working), you might do a blocking experiment using the same antibody but unlabeled to compete with labeled antibody binding, or use isotype control approach when you add a labeled antibody of the same isotype but different specificity to show lack of binding in that case.

Good points! Just wanted to add the possibility to do a blocking experiment by pre-incubating the antibody with the peptide that was used to produce the antibody and to use that as a negative control. In that way you will get a more accurate cut-off value, for example if one type of cell is more "sticky" than the other one. Also, check whether you have big differences in signals between/among cancer and normal cells without the antibody. This will of course affect your cut-off. I agree that isotype controls are not always a good alternative (I couldn't see that Eiji Miyoshi suggested that) especially if you are to compare results obtained by different lot numbers, which may show variations in the ratio antibody:fluorochrome. However, if you always stain normal and cancer cells in parallel and compare them for each experiment (might be necessary to normalize results) this shouldn't be any problem. Finally: whatever results you obtain, I would have confirmed them also with other methods, for example immunostaining for microscopy and western blot. As Anna C Belkina points out, with your approach it's not possible to conclude that your antibody binding is specific. To do that it would be necessary to block or to reduce the expression of the protein.

Now that I re-read it, it is true that prof. Miyoshi suggests an indirect staining. My apologies for the mistake. I understood that he recommends using a secondary antibody as an isotype control.

sure you could. 

Thank you for everyone, and I have additional question, when I wanna use FITC conjugated antibody as secondary Ab, then how can I avoid the non specific signaling? I have an idea as below, Firstly, I would stain the cells with anti-protein XX specific ab with biotinylated conjugation, and next, I wanna stain the cells with streptavidin with FITC conjugation. Then, I want to use cells were stained with just FITC conjugated streptavidin as the negative control. Can I do my experiment as above design? 

Yes, but you could also consider to pre-incubate the primary Ab with a blocking peptide as a negative control. Remember to titer your secondary antibody or FITC conjugated streptavidin well, just like the primary antibody.

Thank you so much~! Now, I totally know how to do. Thank you~

Yes Jin! u need to keep unstained  for each type of cell  and antibody as negative control

Hi all,

I see splendid Q&A here, but are KO cells lack of corresponding surface protein better as biological ctrl when possible?

Same type of sample (eg. cells) in which the level of the protein of interest has been reduced, or even been brought to zero, is a good negative control. Of course, there is always a risk that your sample (eg. cells) has been modified by the reduction of the protein, but it is always a good negative control. Always best to compare a result with other techniques, when possible!

When we use FACS to sort or to analyse cell cycle in our lab, we definitely use a population of unstained cells as a proper negative control. Also, when I analyse cell cycle, I use a population of PI-stained cells without any inhibitor treatment. It gives me a proper positive control for cell staining.

All of the above advises by other researchers and correct and helpful. Thanks.