Is there an appropriate way to normalize two P(r) functions on data collected at different times on different detectors and at different concentrations?

Everything looks good, but when I generate the P(r) plots, they are not scaled to each other at all. The two samples are of the same multi-domain protein in different configurational shapes. Normalizing to a peak maximum is inappropriate because that is exactly what we expect to be different in the samples. My initial thought was to integrate both to an area of 1, but that doesn't work because Primus automatically selects a number of XY data points for the P(r) of each sample. Graphing programs just add up the points for the integral, and since the number of points are different, I still can't scale to 1 appropriately.

There must be a simple solution to this issue.

## Normalizing P(r)

### Re: Normalizing P(r)

typically, you have to normalize P(R) function to unity by dividing by max value of P(R) function. This way you

can compare different P(R) functions.

can compare different P(R) functions.

### Re: Normalizing P(r)

Assuming the two curves come from the same protein in the same oligomeric state (but different conformations) they would have (almost) identical I(0) values on an absolute scale. So if you normalise the experimental data to the same I(0) before distance distribution analysis your p(r) functions will have the same area.