Dear amber users:
We ran mds on a modified oligo in explicit solvent under sander_classic and sander in amber6 and sander in amber5. The runs went for 2 ns. After about 1 ns, the total energy of the system started to decrease and the rate of the decrease appears to be accelerating as the run continues. There is not as much of a decrease using sander in amber6 but it is nonetheless noticeable. Other quantities (temp, pressure) are constant but the density is increasing with corresponding decrease in volume during all three simulations. We don't really believe that the decrease in total energy is real and would like to know it's source and what to do about it. We have posed this question to the reflector in the past and received the following advice from several sources: change ntcm to 1 and nscm to 10000 in order to to remove the COM motion and rotation about the COM. According to the manual the nscm flag is ignored for both belly and periodic simulations. Since the decrease in energy doesn't start until about 1 ns, is it
enouogh to set ntcm to 1?
Thanks. Sue
Received on Sat Sep 08 2001 - 15:00:25 PDT
