I plotted a concentration -response curve using raw values, followed by linear regression ( y=mx+c) to calculate the ic50, as this is how it is done in literature ( s least for work related to cell culture. Is this a valid method? and if yes can someone validate this method with a peer review paper or book or link?
Any assistance is greatly appreciated.
I wouldn't trust a simple linear regression for a dose-response curve as they are typically sigmoidal in nature. Graphpad (or other statistical software package) ought have the equation built it (it's kinda complicated). The link below works on-line (requires Java)
http://www.changbioscience.com/stat/ec50.html
I would suggest you to use GraphPad PRISM sofwtare. Effective concentration (IC50) values can be calculated by regression analysis, using the dose-response curves generated from the experimental data. then data can be presented as sigmoidal curve.
Plot % inhibition versus concentration. Use a nonlinear least-squares regression program to fit the data to a suitable equation, such as this one:
% inhibition = 100 x [I]/(IC50 + [I]), where [I] is the inhibitor concentration.
This is the simplest equation and a good place to start. There are variations that allow for cooperativity of inhibitor binding and incomplete inhibition.
It is helpful to plot log10 of the inhibitor concentration on the x-axis to spread out the graph. When you do that, you should see a so-called S-shaped curve.
This is ok. Alternatively you can use Matlab software to get the perfect curve fit, and it will also give you the equation.
- Badges
- Science method