I have done a competitive ELISA for the first time and I am not sure about the way to calculate the concentrations? I did a log transformation of the data to generate the standard curve? Is this better than the 4 parameter logistics? What is the difference?
First, you should have a high value in wells with the antibody added but no competing antigen (or vice versa, as these assays can be done with either antigen or antibody coated on the well). I would generally subtract the background values (obtained in wells with only the competing antigen added, no primary antibody, but with secondary antibody). Then, plot the data in a variety of ways (linear, log-linear, log2-linear, log-log, etc and compare R2 values. In general curve fitting methods that give higher R2 indicates that this method of curve fitting is better. If that happens to be 4 parameter logistic, then use that one but if a similar r2 value is obtained with a simple linear plot, I suggest using the simpler plot. If you have software such as Prism 6.0, you can select from a variety of linear and non-linear curve fitting methods. This software will also give you the formula for the line, and you can use that formula and the value obtained with the unknown to calculate the concentration in the unknown based on its absorbance. You can also set up the software to solve the equation for you and automatically show values for unknowns on a results page. There are many software packages that will let you do the same thing. I like Prism more than more powerful programs such as SAS because Prism is very user friendly and does not require one to learn any non-intuitive data entry conventions or learning a scripting language. A word of caution-the software will give you a number, even if your unknown value is outside the range of the standards. Such values can be wildly incorrect, and they shouldn't be used as anything but rough estimates.
Hi Mohammed,
you can only derive a protein concentration from a competitive ELISA under very special circumstances. If you are not sure whether you have these conditions in your experiment you should work with IC50 values. You know that you appropriate conditions when you are able to derive the analytical equations from the mass-of-law action. A few rules of thumb: First, you need to be sure that the concentration is much higher than the equilibrium dissociation constant, if not you are making a KD determination or a mix of both KD and C. Second, you have to see a propor (i.e. S-shaped) competition curve. Third, monovalent soluble Ligands / antibody fragments are best, bivalent monoklonal Antibodies are already problematic (see e.g. "Avidity" and bivalent binding to the coated antigen) and polyklonal Abs even more complicated. In such cases you can only derive a concentration by comparisson with a known standard, in this case you simply use the fold difference in the IC50 values for calculation. As pointed out by Stephen, there are many pitfalls, so be sure to make enough controls, never trust the software and don overinterprete your results.
Hi Mohammed,
Please see a paper by Peter Kuzmic and myself on fluorescent displacement assays. One can use the same program we used to determine the KD of a competitive ligand binding to a protein and displacing another ligand. I don't remember the details, but the methodology is in the paper. And you can work above, below, or at the KD in terms of ligand and protein or receptor concentrations. The program we used takes into consideration conditions when the receptor concentration and ligand concentrations are above or at the Kd for binding. I think the program today that can be used is Dynafit.
- Badges
- Science method
- Similar topics
- Medicine
- Immunology
- Immunology Techniques
- Immunoassay
- ELISA