If docking is carried out using AutoDock Vina, then affinity values [in kcal/mole] is obtained. Now if I want to obtain Kd [binding constant], then how can this affinity result from autodock can be utilized to get Kd values?
It is however very unlikely that you would find any significant correlation between the predicted and calculated Ki's. Usually, the performance of the AutoDock score (and any other docking scores) is measured in terms of success rates, i.e., what fraction of predicted actives are actually active. Another useful performance metrics is sensitivity, i.e., what fraction of actives has been predicted as actives.
Kd (kcal/mol) = e^y where y=(delta G/-1.419) for T = 37 Celsius
When I ran this calculation for my own dockings assuming 37 Celsius conditions, I found that a 1 integer unit difference in delta G is about a 2 fold difference. So docking at delta G = -13 binds ~ 68 fold better than a molecule docking at delta G = -7.
I went back and redid my calculations. T in Kelvin is 310K for 37 Celsius.... my newest math suggests about a five fold change in Kd for each integer unit of delta G. This means predicted affinity difference between delta G = -8 and delta G = -10 is 5*5 =25 fold. This matches experimental values better than my previous value of 2 fold. It equates to microMolar potency at Delta G = -8 and tens of nanoMolar potency at Delta G = -10 ... ignoring influence of PK and PD completely lol. >Matt
As you have said earlier that you have taken 310 K temp. during the calculation of Kd. Is 310 K is the default temperature of Autodock? On what basis you have taken the Temperature?
I performed autodock vina and obtained the binding affinities, e.g. -10 kcal/mole. Hence, -10=-RTlnKd, after solving Kd (at 298K), I get 2.2x10^7 M. Is this correct? This means Ka is 5.5x10^-8. Is there any reference which support it? Thanks in advance.