Gabriela de Lima Menezes Madam, you can try first 300K !!!

Mohd. Babu Khan I’ve already performed a MD simulation at 300K, but some contributor argued that this temperature is for mosquito and 310 K is for the human host (my target). It’s a protein from Alphavirus, an arbovirus. Do you think the force field is good enough to capture this 10 K difference? I’m using AMBER ff99sb-ILDN.

How can I argue 300 K is ok?

Thanks!!

If the in vitro experiments are carried out at room temperature, run MD simulations at room temperature. If the in vitro experiments are carried out at normal human body temperature or if there are no complementary experiments, run MD simulations at normal human body temperature. And make sure that all the MD simulations in the set are carried out at the same temperature so that the structures and energies from different MD simulations in the set can be collected into a single ensemble for analysis.

Thanks, Martin Klvana ! and what about gen-temp and ref-temp? they should be the same? Or in case of human temperature, should I use 310 K for ref-temp and always use 300 K for gen-temp or 310 K as well?

i don't use Gromacs so correct me if i'm wrong, but gen-temp is the initial temperature, isn't it? The initial temperature is set only once: for the first stage of the equilibration MD simulation and it should be very low, e.g. 5 K. In further stages of the equilibration MD simulation, as well as in subsequent production MD simulation, gen-temp is not defined: the current temperature is simply determined from the atomic velocities, and so only the ref-temp parameter is defined.

310 K is normally used as human body tempereature. I guess you are not doing anealing, so all ref should set as 310 K. gen_temp option will come only when you are doing simulated anealing or you intend to generate temperature instead of velocity.

Hi,

Though, 310K is human body temperature, however, in literature, there is countless examples where people used 300K and inferred results. and I do not think AMBERff99SB-ILDN can capture this effect. And as a matter of fact, this is very small temp difference so most MM force fields may not truly capture this.

Other factors may also be in play as the NVT vs NPT ensemble, protein size and secondary structure contents etc.

The easier solution is to compare if you have computational resources, otherwise, a crude argument is that, ff99SB-ILDN has been parameterized against 300K (I hope, I read this paper but do not remember exactly).

If you are doing regular simulation (not simulated annealing), then gen-temp is not required, just set "gen-vel = yes" for NVT equilibration [and gen-vel=no for subsequent simulations, NPT equilibration and production run], and it will set the velocity (and hence temp) based on your "ref-t".

Good luck