I have 1215 peptides between 4 and 13 mer in length. I want to dock them with peptides using Autodock Vina, but I used AMBER99SB instead of Kollman charge to give better and reliable results. I did not get low binding when I expected low binding, I got a binding score between -7 and -7.5 in only 26 samples of 1215. However, I was not sure if it was more correct to do it by loading Kollman charge?
Hi.
The charge scheme selection is tricky as there is no one valid definition for charges.
When talking about docking scores, remember that, for some programs, the lower the score (more negative), the better the docking, i.e., better binding.
As you can use charges calculated for different schemes, you can always compare your results within each scheme and not between them. A good exercise is to do docking calculations with different charge schemes for complexes with already known binding affinities.
Thank you for your response. Based on my experience, I have observed that using AMBER99SB yields better results compared to Kollman charges, with a difference of approximately 0.2-0.3 in binding scores. However, the computation time increases by about 10 times. Additionally, the `andbranch` and `torsdof` values are different, as AMBER99SB performs more detailed and in-depth calculations.
Therefore, while AMBER99SB offers more reliable and accurate results, it comes with the trade-off of significantly longer computation times.
Best regards,
Mehmet