Dear Raquel Sánchez Gutiérrez,

Before calculating the SI index, I suppose you have to check the positive frequency for each dilution.

For example, If 5 ul gives the positive frequency of 10 %, then you can only choose those dilutions that gives the positive frequencies of 10 %. As such, if 1.25ul gives 10 % and 0.625 ul gives 7 %, then your choice should be 1.25 ul or higher and never 0.625 ul.

So please clarify whether you have confirmed the positive frequencies before calculating SIs.

Dear Raquel,

Also, it would be useful to look in the literature what are the expected percentages of positive cells for each marker.

From your last two pseudocolor plots, it appears that your markers are expressed in low percentages, this would explain the shape of your graphs. So, you need a sample that express in higher proportions the molecules that you are analyzing.

Are you using cell lines, or biological samples such as PBMC, splenocytes? Did you discriminate necrotic/dead cells to exclude non-specific staining?

I hope this helps.

Heriberto

Dear Raquel,

When titrating an ab your goal is to choose the biggest dilution possible (to save money) with the best separation between negative and positive population. You can do this with calculating the SI but depending on the stainings it can be obvious to the naked eye. I should say I’ve done tones of titration in my life for flow cytometry and never calculated this value. Nevertheless, in order to do so you should:

1. Know the ab you are using, meaning know the tissue/cells that express it the most. This means that even if you intend for the ab for a new tissue/cell you should try to use the “best tissue/cells” as a positive control;

2. Have the same % of – and + cells in the different dilutions. Only this way you will see a shift in the separation, otherwise you might be looking at no separation between +and - in the highest dilutions.

3. Besides excluding dead and dublets when analysing you should also pay attention to number of events. If your + population is very low in % you might have to acquire more events to trust the plots.

Think more about the purpose of what you are doing and less about a number given by a programme.

Best of luck.

Silvia