CHARMM Development Project
Posted By: ckc1117 size of solvation box - 09/22/13 03:56 AM
recently, I performed MD for alanine dipeptide in water box. The population of the backbone dihedral angles (phi/psi) were quite different in water boxes with different size. we can fine PPII, alpha-R, alpha-L in cubic water box with its sidelength of 28, but in the box with its sidelength of 42, there were no population for alpha-L.
the force field are the same, and the CMAP is considered.
my question is, how did the box size affect the population of the backbone dihedral angles, and how to decide the most probable size of the solvation box?
thank you so much for your time!
Posted By: lennart Re: size of solvation box - 09/22/13 08:33 AM
It is unlikely that 28A or 42A side length would make a drastic difference for the alanine dipeptide. Are your results converged? There are tons of literature using this model system, you can try to reproduce some of it.
A series of converged simulations with increasing side lengths will tell you what is needed in order not to have a size dependence.
Posted By: ckc1117 Re: size of solvation box - 09/24/13 09:17 AM
Thank you so much for your time, I'll check the literature and try to reproduce some of it.
ps: I've try 30 A and 42 A side length for ALAD in CHCl3, the dihedral distribution shows no big difference, but the MD of ALAD in 34 and 42A DMSO boxes are quite different.
Posted By: lennart Re: size of solvation box - 09/24/13 03:44 PM
The fundamental question is whether your simulations are converged; this must be established before your start making comparisons.
Posted By: rmv Re: size of solvation box - 09/24/13 04:10 PM
Running replicate simulations, esp. starting from different molecular configurations, is one way to try to answer this question.
Posted By: ckc1117 Re: size of solvation box - 09/25/13 01:44 AM
Dear Prof. Lennart, I am sure the energy, the temperature, the presure and etc. are convered.
Posted By: ckc1117 Re: size of solvation box - 09/25/13 01:56 AM
Thank you for your advice, I'll try it with other initio structure (the origin ALAD is extended).
Posted By: lennart Re: size of solvation box - 09/25/13 05:09 AM
Yes, but is the conformational distribution converged?
Posted By: ckc1117 Re: size of solvation box - 09/26/13 10:12 AM
Dear sir, I am sorry that I don't know how to make sure the conformational distribution was converged, does it have another criterion ?
Posted By: lennart Re: size of solvation box - 09/26/13 10:39 AM
AS Rick said, run several independent simulations and see if your distributions are the same; if you have a long simulation you can see if the distributions obtained using different parts of the trajectory are the same. Ideally you should have multiple transitions between the accessible conformations.
Posted By: ckc1117 Re: size of solvation box - 09/27/13 02:13 AM
Thank you sir, the MD with different initio configuration is running, I'll let you know if there is any progress. Thank you so much for your time! ^o^
Posted By: ckc1117 Re: size of solvation box - 10/03/13 07:07 AM
Dear all, replicate simulations were performed for ALAD in water box, the backbone dihedral distributions are different. There are populations in the alpha-L region with initio-structure of alpha-R; while this population can not be observed with the initio-structure of alpha-L.
So, how can I be sure that the conformational distribution is converged, is it necessary to increase the production time of MD simulation ?
Posted By: Yoda Re: size of solvation box - 10/04/13 01:05 AM
Different distributions mean simulation converged is not.
Posted By: Kenno Re: size of solvation box - 10/04/13 06:13 PM
LOL that's funny! laugh
Posted By: ckc1117 Re: size of solvation box - 10/06/13 06:27 AM
In my opinion, the conformational distribution would be different from the converged MD simulation with varied initio structure.
Posted By: lennart Re: size of solvation box - 10/06/13 08:35 AM
I'm not sure I understand. Are you saying that converged MD simulations will give different distributions if they are started from different conformations? If this is what happens, the MD simulations are not converged.
Posted By: ckc1117 Re: size of solvation box - 10/06/13 03:40 PM
So if we perform the MD simulation for a polypeptide with 40 residues, and start with beta-sheet or alpha helix, can we get the same conformational distribution? or if the production time is long enough, the conformation distribution would be the same too ?
Posted By: Kenno Re: size of solvation box - 10/06/13 04:28 PM
Yes, absolutely. (Note, however, that for some problems, "long enough" may surpass anything computationally or even experimentally feasible.) Equilibrium only depends on the energy profile, not on where on that profile you start off. The relationships between energy profile, equilibrium and kinetics are described in basic chemistry textbooks - that fact that you had to ask that question indicates you should try one.

On a related note, fundamentally spoken, the term "converged" is relative; one could argue it should be "converged with respect to free energy barriers of magnitude X" or "the sampling of conformations Y and Z is converged" (where Y and Z would be at different sides of barrier X). For example, if one starts a protein simulation from a folded state, one usually will not expect the sampling of all possible misfolded states relative to this folded state to converge; the time scales on which this converges are often too long to be directly accessible even experimentally (especially in the absence of chaperones). At the same time, one will often want the sampling of certain side chain conformations to be converged.

Practically spoken, we commonly use the term "converged" in the sense of "converged for our purpose". For instance, if your purpose is to study the (φ,ψ) conformational distribution of the alanine dipeptide, and you get significantly different (φ,ψ) distributions depending on the starting conformation, then your simulations are clearly not converged for your purpose. Luckily, for such a simple system, convergence should be computationally very feasible.
Posted By: lennart Re: size of solvation box - 10/06/13 05:31 PM
If you cannot, it becomes pretty meaningless to compare distributions obtained with boxes of different size, which is where this thread started. Sigh. Discussion with the supervisor is worth a thousand forum-posts.
Posted By: ckc1117 Re: size of solvation box - 10/06/13 11:10 PM
Thank you so much for your time! I am really appreciate your help!
I'll check the RMSD and find out a way to make sure the system is converged.
I wish I have a supervisor to discuss, however, I am on my own. But your advice are very helpful for my research work, thank you!
Posted By: Kenno Re: size of solvation box - 10/07/13 02:49 AM
You must be employed by someone. You mean you have a supervisor but they know nothing about this field? Or you're doing this at home as a hobby?
Posted By: ckc1117 Re: size of solvation box - 10/07/13 05:13 AM
I got a Ph.D degree from my boss who's major is vibrational spectroscopy 3 years ago, and I'm running my own lab in a university. So right now, I am my supervisor.
Posted By: Kenno Re: size of solvation box - 10/07/13 06:18 PM
Sorry for the rant ahead, but it bothers me that there are so many physics, organic chemistry and molecular biology PIs who decide they want to do a little bit of "molecular modeling" at the side. This might have been possible 30 years ago, when the field was in its infancy, but since then, it has become a branch of science in its own right. A molecular biologist is not going to buy a set of glassware, evaporators, IR spectrometer etc., give them to a grad student that had good grades in chemistry and expect them to synthesize molecules. Neither is an organic chemist going to buy a PCR machine, equipment for southern blot,... and expect a grad student to be doing molecular genetics with them. What they're going to do if they want to diversify is either hire one or more senior postdocs with longstanding experience in the desired field, or go spend several months in an expert's lab themselves. Why do people think they can just buy a few computers, download software, and expect a person with little experience in this field to perform solid computational chemistry or biophysics with it?

Since I started here, we've had 2 young faculty members in the field spending several months in our lab. Afterwards, they were happy with the things they learned. Maybe you should consider doing something like that?
Posted By: lennart Re: size of solvation box - 10/07/13 08:06 PM
Kenno is unfortunately right - this is a fairly difficult field, with many complications and pitfalls. Much more difficult than e.g. quantum chemistry.
Posted By: ckc1117 Re: size of solvation box - 10/10/13 01:40 AM
Yes, you are both right, fundamentals are the building blocks of fun. I was considering to become a postdoc (maybe too late now, I got Ph.D 3 years ago) or a research fellow in this field, but I am doubt that any group would like to take somebody in just for learning but not producing datas right away ?
maybe, I should prepare a CV to see if there is a chance for the improvement of MD skill.
Posted By: ckc1117 Re: size of solvation box - 10/10/13 03:11 AM
The Nobel Prize in Chemistry 2013 was awarded jointly to Martin Karplus, Michael Levitt and Arieh Warshel "for the development of multiscale models for complex chemical systems".
Posted By: Kenno Re: size of solvation box - 10/10/13 03:40 AM
Originally Posted By: ckc1117
Yes, you are both right, fundamentals are the building blocks of fun. I was considering to become a postdoc (maybe too late now, I got Ph.D 3 years ago) or a research fellow in this field, but I am doubt that any group would like to take somebody in just for learning but not producing datas right away ?
A lot of institutions and granting agencies have small grants that allow beginning faculty to visit another group for a few months to focus on learning new techniques. That's what the 2 people I talked about earlier did. You'd be bringing your own money and you'd be learning by doing, with a fair chance of a joint publication. This should make it moderately attractive for someone to host you.

Alternatively, channels like the CCL are overflowing with postdoc openings. Since you have experience beyond being a PhD student and you're already running a lab, I'd imagine you'd make a very competitive candidate; independent research skills are highly valued in postdocs. But of course, that would be a big change in direction for you.
Posted By: ckc1117 Re: size of solvation box - 10/10/13 05:57 AM
Thank you kenno, I would think about it!
PS: is there any posible position in Prof. Alex's group ?
Posted By: Kenno Re: size of solvation box - 10/10/13 03:39 PM
I don't know, you'd have to ask him. Right at this moment, I'm a bit doubtful, but generally spoken, people come and go and positions open and close. If he can't host you, there are plenty of other experts to try.
Posted By: Hyuuga Kawai Re: size of solvation box - 10/13/13 02:18 PM
Originally Posted By: ckc1117
The Nobel Prize in Chemistry 2013 was awarded jointly to Martin Karplus, Michael Levitt and Arieh Warshel "for the development of multiscale models for complex chemical systems".


Martin Karplus is a theoretical chemist. He understands and predicts chemical processes..
Posted By: Kenno Re: size of solvation box - 10/13/13 08:46 PM
Welcome to the official CHARMM website, home of the CHARMM development community. As you might be able to infer from the above links, most of us have a pretty good idea of who Martin Karplus is...
© CHARMM forums