<p>Actually i know people using Go-models to study protein folding for proteins as large as 300 aa. Now, in these cases you need to use biasing potentials and not temperature exchages (maybe both?). In principle, the REMD will help you with the comformational exchange but it also includes a new level of complexity since the temperature distribution for the correct exchange ratios is pretty much an empiric issue. In principle, there is no problem to run these simulations independently and then combining them by WHAM taking the temperature as the only bias. Just consider the parallel tempering as a sampling enhancer the same way as it is used also with hamiltonian hopping.</p>
<p> </p>
<p>If you go for the umbrella sampling way you can use the PLUMED plugin for colective variables. There you can apply a bias directly to the number of contacts of the structure (a contact map bias) or maybe to the gyration radius. In principle, once you unbias your histograms you can calculate the free energy profile of any colective variable using a "multidimensional histogram" trick. In the end, what you need to now is how much energy you are adding to sample that state. </p>
<p> </p>
<p>Once you have the free energy profile at the folding temperature (this also needs some trial and error) you should be able to measure the height of the energetic barrier between states which is in principle what determines that transition rates. </p>
<p> </p>
<p>As mark said, there are tons of papers of this as "Go-models" ar practically the only way people had to explore the protein folding thermodynamics in the computer.</p>
<p> </p>
<p>regards</p>
<p> </p>
<p>Felipe</p>
<blockquote>----Mensaje original----<br />De: Mark.Abraham@anu.edu.au<br />Fecha: 03-ene-2012 22:52<br />Para: "Discussion list for GROMACS users"<gmx-users@gromacs.org><br />Asunto: Re: [gmx-users] Re: Folding rate<br /><br />On 4/01/2012 12:35 PM, bharat gupta wrote:<br />> Thanks for all your replies. I want to know this can be done in <br />> gromacs or not - using REMD with structure based models generated from <br />> SMOG server to study protein folding and unfolding ??.<br /><br />Well, it can be done, but you probably don't have enough computer to <br />fold a 230 residue protein at atomistic resolution (or maybe even <br />coarse-grained).<br /><br />> Also, I have a question about how to determine the exchange <br />> probablities for a particular REMD experiment and also how many <br />> replicas do we need to consider, does that depend on the temperature <br />> list generated from the T_REMD server??<br /><br />There's a significant literature on these subjects. I suggest you read <br />some of it. Short answer: pick the highest temperature according to the <br />size of the largest barrier you expect to cross (good luck guessing <br />that), have around 20% exchange acceptance, and be prepared to observe <br />where the replica-flow bottle necks are and to iteratively refine you <br />temperatures.<br /><br />Mark<br />-- <br />gmx-users mailing list gmx-users@gromacs.org<br />http://lists.gromacs.org/mailman/listinfo/gmx-users<br />Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!<br />Please don't post (un)subscribe requests to the list. Use the <br />www interface or send it to gmx-users-request@gromacs.org.<br />Can't post? Read http://www.gromacs.org/Support/Mailing_Lists<br /><br /></blockquote>
<p> </p>