Dear APBS users,

I am writing to ask you a question on how long my MD simulations should be to obtain reliable data for protein/peptide interactions.

In general, it is my understanding that MD simulations shorter than 500 ps were sufficient for MM/PBSA results (Kuhn et al., J Med Chem, 2005; Weis et al., J Med Chem, 2006). These works were performed on ligands.
Notably, the first MM/PBSA works on protein/peptide interactions (I am referring to Massova and Kollman, 1999, and Huo et al., 2002) presented very short MD simulations. I assume that the limitations of the computational costs at that time were a major issue for the MD simulation length, but still the results worked out pretty well.
Also in some recent papers investigating a protein/protein interaction (I am thinking of Moreira et al., Theor Chem Account, 2008, or Majmudar et al., Proteins, 2011 -- the latter starting from a docking model) the production runs are not extensively long (5 ns and 0.4 ns, respectively).

I understand that -- hopefully -- the longer the simulation, the more reliable the data, but I would like to have your opinion about it.

Is there any difference between ligands' and proteins' (or peptides') results?
Can somebody indicate me whether someone has ever pointed this issue?

Thank you very much in advance.
Best regards


Dimitrios Spiliotopoulos

Dulbecco Telethon Institute c/o DIBIT Scientific Institute
Biomolecular NMR Laboratory, Basilica 3A2
Via Olgettina 58, 20132 Milano (Italy)
Tel     : 0039-0226434922
Fax    : 0039-0226434153
Email :;
Skype: dimitris3.16
Site    :