Hello, I usually use default medium or, rarely, long parameters for my compounds + known substrate/inhibitor for the adequate comparison. I think the approach is more simple vs attempts to find out some best docking parameters (in my oppinion they depends strongly from the nature of the ligand and protein target). Concerning rotation of amide bonds, I usually mark it as unrotable, but if necessary I may allow rotation of some to compare the results. Good luck.

thank you sir for proper information

You can use the defalt parameters. But If you want to see the best conformation along with the possibility of it to stay at that pose then increase the runs. Usually it will be 100, but can increased upto 250. With this you will have more confidence of the binding pose than doing with 10 or 100. Since it works with Genetic algorithm, you can't expect the same pose and energy if you run the complete simulation again. So, if you are getting -4.36 kcal/mol, it will be +/- 0.5 kcal/mol as i suppose with my past experience.

Hi Ravi,

As Yaroslav suggested, the question is more relavant to comparitive studies with known inhibitors.

Low docking score doesnt reflect that docking prediction is poor if known substrate/inhibitor has low score. So dont worry about docking score values. Try to correlate docking studies with known inhibitors.

I prefer to go with default parameters for ligand preparation.

Reg. Protein preparation, I prefer to 'add all hydrogens' then 'merge non-polar hydrogens' to retain polar hydrogens. I found this method give substantial results than just adding polar hydrogens in practice, not therotically.

Reg. Search and LGA parameters, long parameters are always biologically significant than default or medium parameters. But it is computationally (time, CPU) intense/expensive.

If you are carrying out virtual screening, perform with 'medium paramenters', filter the best ligands then perform docking with 'long parameters'.  If you are working with limited/small no. of molecules, perform with 'long parameters'.

Parameters for limited small molecules, I have used in my research:

LGA runs: 100 (250 when results are not biologically significant)

Population size: 300 (for better sampling of 3D space)

Max. no of evaluations: 1000000 (max. 3000000 when results are not biologically significant)

Max. no of generations: 27000 (While using long evaluations, increasing max. no of generations doesnt matter).

Please check out my articles:

http://www.sciencedirect.com/science/article/pii/S1093326314000102

http://www.sciencedirect.com/science/article/pii/S0040403912009227 (Check supplementary data for detailed methodology).

Hope it helps.

 -- P. Saravanan

P. S.: Please check the other docking related posts in ResearchGate & in AutoDock forums and FAQ of AutoDock.