Dear Stuart, to get a reliable value of IC50, you should assay at least two concentrations above, and two below, the IC50 you expect.  If you're thinking that your coumpond has a 50% activity near 2.5 ug/mL, you might try with 10, 5, 2.5, 1.25 and 0.625.  Remerber to add a positive control group in your assay, it means a group exposed to a drug with known activity.  Also remember that the activity is not linear with the concentration, they are related in a log way (almost always), so, to get a good IC50 value, you must plot the activity versus the log of concentration.

All the best,

Rodrigo

Dear Stuart,

You can use the two-fold serial dilution method to maintain as your mentioned concentration between 2.5 to 200 µg/ml for determination of IC50 value. Here I am sending a article as attachment in which MIC value has been determined by using two-fold dilution method. You can customize it as according to your need. 

Thanks very much

What I am unclear about is why the papers say the organisms  were incubated in different concentrations (2.5-200 µg/ml) of the substances they are testing. Using a 2 fold series dilution it would go 200 µg/ml, 100 µg/ml, 50 µg/ml, 25 µg/ml, 12.5 µg/ml, 6.25 µg/ml, 3.125 µg/ml, 1.5625 µg/ml. So where do they get 2.5 µg/ml. Sorry I know this is basic but this is not my field!

Dear Stuart, to get a reliable value of IC50, you should assay at least two concentrations above, and two below, the IC50 you expect.  If you're thinking that your coumpond has a 50% activity near 2.5 ug/mL, you might try with 10, 5, 2.5, 1.25 and 0.625.  Remerber to add a positive control group in your assay, it means a group exposed to a drug with known activity.  Also remember that the activity is not linear with the concentration, they are related in a log way (almost always), so, to get a good IC50 value, you must plot the activity versus the log of concentration.

All the best,

Rodrigo

Dear Stuart,

I am agreed with Dr. Lopez-Munoz. I think you have got your answer. 

No, you do not have the right answer.

You have to have at least one concentration that does not have any effect at all, two concentrations that produce maximal effect and two concentrations between, that will probably define your strait line between no effect and maximal effect. This is this section of the sigmoid curve that will determine your IC50. It is possible to produce some results which will be fitted to a sigmoid curve with even 4 concentrations (presumed maximal, no effect and two between them) but I would not recommend this.

If the analysis is a problem:

The analysis of the graded dose-response (extrapolation of the IC50 or ED50) is done by means of non-linear regression using the Hill–Langmuir equation which may be found implemented in GraphPad Prism (GraphPad Software, San Diego, CA, USA). The results are most often given on a p-scale, as the mean of –log10 EC50 values experimentally obtained.

I recommend using the method of Lew and Angus (Lew and Angus (Trends Pharmacol. Sci., 16:328-337, 1995), and

Lew MJ1, Angus JA. An improved method for analysis of competitive agonist/antagonist interactions by non-linear regression. Ann N Y Acad Sci. 1997 May 30;812:179-81.

The best book that I know is:

Terry P. Kenakin, Ph.D.:A Pharmacology Primer . There is a fourth edition available. Earlier editions that explain the method correctly may be found online, may be even free. Some chapters are given as separate PDFs

http://www.sciencedirect.com/science/book/9780124076631

Thanks very much for all your input. The point that confused me was that so many papers on the organism I am working on state they tested the substance at 2.5-200ug/ml. Using a 2 fold series dilution from 200ug/ml would go to 1.5625ug/ml so I can not work out how most authors seem to start at a concentration of 2.5ug/ml. I have been told not to worry about as the range of concentrations I am using will cover it! Thanks again for all your input, very useful 

From the previous papers or your pilot study you may learn how low concentration will “just” not have any effect and how high concentrations will produce “just” maximal response. To make sure that you know which concentrations really produce maximal response, you better always have two quite close concentrations that produce maximal response.

Of course, if your response, for each concentration, are reached rapidly and there is no time effect, you should have as more points as you can have between those maximal and minimal, may be 6, 8 or even 10 points (by using that many concentrations).