You are here

Archive topic 2 (transferred from old site)


How to choose minimum and maximum temperatures for REMD

Sun, 03/17/2013 - 04:43 — PengJH

Dear Developers,
I have notice that UNRES/REMD is mostly used for folding essays. Is it proper to do REMD with large system for better sampling of conformational changes using UNRES? And what is the main factor to determine the upper and lower temperatures for the REMD systems?


Sat, 03/30/2013 - 07:27 — adam.liwo

Re: How to choose minimum and maximum temperatures for REMD

Hi Peng,

Yes, you can use UNRES/MREMD for large systems to study conformational changes; we recently did this to study Hsp70 chaperone motion (Golas et al.. JCTC, 2012, 8, 1750). In such cases it is appropriate to start MREMD from experimental structures. The problem might be the convergence which is always more severe with larger systems. I'd suggest you to monitor the convergence of MREMD (e.g., the convergence of the temeprature dependence of an order parameter such as radius of gyration). To do do, you compare the results from consecutive time windows; the WHAM utility enables you to compute the averages from several windows automatically.

In most cases, it is OK to set 250 K as the lower and 500 K as the upper bound of temperature. The rationale is to cover the range of the transitions the molecule undergoes. If you are experimenting with a new/modified force field, you might need to change the boundaries. Typically, you should use at least 16 temperatures in this temperature range and the spacing between consecutive temperatures should be denser in a region of transitions (see U.H.E. Hansmann's papers on this).