Minimum Data Points for elisa standard curve - (Jan/29/2015 )

Hi

I have two questions:

1. In my Elisa results, I have 4 standard calibrators and 1 negative standard (conc= 0 IU/ml).Is it practical to use 4PL or 5PL equations for only 5 data points( 5 standards)?
 
2. In Elisa test instruction that the company gave me, it is noted that the calibration curve should be a linear plot. But my results can be better explained by non linear curves. Am I allowed to disregard test instruction in this case?

Any help would be appreciated.

The choice will depend on where you will be presenting these results and whether your standards match those of the ‘typical curve’ provided by the company.

 If you are doing the elisa for an external client then deviating from the ‘official’ instructions by changing the calibration curve can cloud your results. It is much easier to say “elisa was performed following the instructions provided” than to validate your changes.

If this is for an internal research thesis you have more freedom to include a paragraph on why you would change the standard curve, contrasting the better fit as you increase the number of parameters against loss of degrees of freedom and the loss of robust error estimation with nonlinearity.