trouble with MC

You successfully prepared your system. Now: How do I tell Hippo to produce a trajectory from it? What parameters to choose? And what should I do when things go wrong.
martin
Developer
Posts: 38
Joined: Sun Jun 15, 2008 10:27 pm

Re: trouble with MC

Post by martin » Tue May 05, 2009 2:38 pm

In protein folding proteins can get stuck for very long times. In particular if the sequence has a lot of charged sidechains and the temperature is too low. I am seeing similar behaviour at the moment with a peptide that is now at 2 billion MC steps and still not folded propertly.

Using a very hydrophobic sequence shows that the program runs fine.

There are several options you could try:
1. Higher temperature (e.g. 350K or 400K) or
2. Replica exchange (REMC), say 20 replicas from 300-500K. The latter worked very well for our systems but you can also try a broader range if your system is very stuck: 300-600K.

You can run REMC on one processor in principle (it's slow though). It generally greatly speeds up the sampling.

Sorry i can't help any better, but trapped states is a very common problem in protein folding, most of our simulations took many trials and months to run.

Good luck!!

Best wishes,
Martin

m.kalavera
Posts: 29
Joined: Mon Feb 09, 2009 10:58 am

Re: trouble with MC

Post by m.kalavera » Wed May 06, 2009 7:18 am

Hello Martin,
thx for your reply at first. In my case the protein is a highly charged one and if you say this is causing the trouble, I am ok with it. So I will run my simulations for a much longer time and I will try the replica exchange method.
After this is done, I will tell you if and how it worked out.
Thx for all your support.
Your m.kalavera

Post Reply