I am doing DPP4 inhibitory studies of phyto constituents at 4 different concentrations (in triplicate) of each constituent through Flourescence method. To plot graph for the determination of IC50, on Y-axis whether we have to take Absorb / % inhibition activity? Kindly provide me step wise procedure in graph pad prism and excel. I used Flourescence based invitro assay kit and used DMSO to solubilise my compounds. (i had taken avg. of Absorbence to determine % inhibition activity of particular conc.)
It doesn't REALLY matter that much. I prefer to convert to %Inhibition (or some other such value) to make the numbers mean something less arbitrary, but unless you screw up your math, the graph should look pretty much identical either way - and whatever algorithm you use to solve for IC50 should give you an identical (or nearly identical) answer.
A common way to make the graph is to put % inhibition on the y-axis and inhibitor concentration on the x-axis. Use nonlinear regression to fit the data to a suitable equation, such as the 4-parameter Hill equation:
% inhibition = (MAX x [I]n)/(IC50 + [I]n),
where MAX is the maximal % inhibition (normally = 100), [I] is the inhibitor concentration, n is the Hill coefficient (normally = 1).
Percent inhibition is calculated as follows: % inhibition = 100 x (1 - (X-MIN)/(MAX-MIN)),
where MIN is the fluorescence measurement with no inhibitor, MIN is the fluorescence measurement with complete inhibition, and X is the fluorescence measurement at the specific inhibitor concentration.
GraphPad Prism has suitable IC50 equations. Excel does not, but they can be found in the Excel add-in XLfit, if you are willing to pay for it. They are also available in Grafit, and there are free on-line sites for making the calculation.
Thanks every one for your responses. below is my problem --
Conc Extract 1 Extract 2
(% inhibition) (% inhibition)
100 55.65572 55.96155
50 43.17258 44.22831
10 22.37284 23.67526
1 10.26158 12.93546
in the case of above values of two extracts, %inhibition may not be having much difference. But when calculated using prism i noticed drastic difference in the IC50. Here is the sequence of steps that i followed.....
new - XY table - checked Y with "enter 2" - create - conc in X column and % inhibition in Y column - analyze - transform the x value (X-logX)-checked create new graph of results - selected four parameter variable from non linear regression - not checked interpolate unknown from std curve and checked least square fit from fitting method
fraction 1 :
Conc. %inhibition
50 (57.5687)
10 (25.86706)
1 (16.00357)
0.1 (18.15835)
0.01 (13.72173)
from these values we can assume IC50 may lie b/w 40-50. But in prism the value has been shown around 20.
Kindly help in this regard.
Prism is most likely assuming this is your entire dose curve. Again: it doesn't know, or care, what the numbers actually are. Since your highest number is 55%, it assumes that the highest number is 55, unless you explicitly tell it otherwise (which you are not doing, using a 4-parameter function). Further, you have a clean plateau at about 15% inhibition: while problematic for a few reasons, that doesn't change that the algorithm will detect it as the lower plateau. Thus, your inhibition range is from 15 to 55 - 40 total units. By your data, the concentration needed to elicit 50% of the maximum effect (ie: 20 units) would be about 20.
To get a value closer to 50, you would want to use a 2- or 3-parameter function, and set the max (and possibly min) to 100 and 0, respectively. But ideally, you'd go back and redo the experiment with higher concentrations of drug: you really want a plateau to tell you what the maximum inhibition actually is. Similarly, you'd go back and use LOWER doses of drug: as SOME point, you should hit 0% inhibition, and if you don't, there's something wrong with your experimental protocol.
in determination of IC50 using prism :
in nonlinear regression analysis - under Dose-response Inhibition, on what basis options has to be selected.
under Dose-response Inhibition, the following options are available:
log (Inhibitor) vs response (three parameters)
log (Inhibitor) vs response -- variable slope (four parameters)
log (Inhibitor) vs normalised response
log (Inhibitor) vs normalised response -- variable slope