I used a Ni-NTA BLI sensor. One protein was immobilized on the sensor. A pool of protein ( out of which one protein binds the immobilized protein ) is then used in the association phase. After that, dissociation is observed in the same buffer.

After looking at data of association and dissociation, it appears that biexponential fit is more suitable for both association and dissociation phases.

---------------------------------------------------------------------

Following are the equations which I have used :

---------------------------------------------------------------------

1. Association

exponential =>    y(x) = a*( 1 - exp(-x/b) )

biexponential => y(x) = a1*( 1 - exp(-x/b1) ) + a2*( 1 - exp(-x/b2) )

2. Dissociation

exponential => y(x) = a*exp(-x/b)

biexponential => y(x) = a1*( exp(-x/b1) ) + a2*( exp(-x/b2) )

I have attached the data points plotted and the best fit of these 4 equations.

---------------------------------------------------------------------------------

Following are the best fit values of different parameters :

---------------------------------------------------------------------------------

1. Association, exponential

a = 0.402 +/- 0.001

b = 68.641 +/- 0.301

2. Association, biexponential

a1 = 0.1600 +/- 0.0001

b1 = 3.0239 +/- 0.0270

a2 = 0.2441 +/- 0.0003

b2 = 73.9547 +/- 0.2147

3. Dissociation, exponential

a = 0.2145 +/- 0.001141

b = 142.361 +/- 1.196

4. Dissociation, biexponential

a1 = 0.1865 +/- 0.0004

b1 = 168.2530 +/- 0.5570

a2 = 0.1624 +/- 0.0014

b2 = 6.0713 +/- 0.0907

It appears to me that indeed there are some low affinity interactions that are taking place and they seem to dominate overall data.

I have uploaded a txt file of raw data points and a graph of data points and fitted-equation.

Note that association is from 630 to 930s while dissociation from 930 to 1230s.

Looking forward for your suggesitons.