new reweighting script

[Pratyush Tiwary]

Hi all

A new reweighting script is now available at this link. It is easy to use for cases with large number of biased/non-biased CVs as well as when there are some CVs experiencing fixed bias. Primarily due to my laziness, the script is in MATLAB, but it should be easy to recode it in any other language if you want. It also might just show up as a part of PLUMED 2 soon, where it could allow you to do reweighting truly on-the-fly. The code follows the paper http://pubs.acs.org/doi/abs/10.1021/jp504920s: Tiwary and Parrinello, J. Phys. Chem. B, Article ASAP (2014) 

On the same location I have made available test data where you can play with the code and fully reproduce the results of this paper, hopefully familiarizing yourself better with the ideas.

It is applicable to both well-tempered and non-tempered metadynamics. It is also applicable to parallel tempering metadynamics. I am not sure if it applies to bias exchange as well, perhaps Alessandro Laio or some other expert can comment.

Some more details:

1. In its current version, it needs PLUMED 1 style output from sum_hills, details in the code.

2. In general, it is an excellent idea to run the reweighting on the biased CV itself, and compare the resulting FES with that from sum_hills. A disagreement between both indicates that the system is far from convergence. In such a case, it is not obvious to me which is better, reweighting or sum_hills. Another clear indicator of poor convergence is plotting the quantity c(t) vs t where t is the simulation time. For converged metadynamics, c(t) should follow c(t) ~ t^gamma (well-tempered). The slope will in general be slightly less than gamma, getting closer as you decrease the width of the gaussians.

3. The code can easily treat cases when some CVs are biased through metadynamics, while some other CVs are experiencing fixed/static bias. 

4. The c(t) expression in this script follows directly from using Eq. 12 in Eq. 3 of the aforementioned JPCB. It is smoother than equivalent Eqs. 13 and Eqs. 14 given in that paper.

Please email me if you have any questions/feedback about using this code. 

Thanks and happy reweighting

Pratyush

--

Pratyush Tiwary

Post-doctoral Researcher, ETH Zurich and USI Lugano

Department of Chemistry and Applied Biosciences

sites.google.com/site/pratyushtiwary/

[Alessandro Laio]

Dear Pratyush, it is a really nice work, and, yes, in my opinion it can be applied also to bias exchange!
Alessandro

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/CAO8zxx4GDwQeNLNWY0ezZu4js_rp9%2B786CgCyom6yeK%2B-ap-Pg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[wyon...@gmail.com]

Hi Pratyush,

Thank you for your sharing the code. Sounds exciting! I am wondering if your code can reweight the simulation with well-temper and adaptive hills? Thanks in advance!

Best regards,

Yong

在 2014年11月11日星期二UTC+1下午3时38分40秒，Pratyush Tiwary写道：

[Pratyush Tiwary]

Dear Yong

I am not so sure if it applies to adaptive-width gaussians, assuming that is what you meant by "adaptive hills" . Especially when the gaussians become very broad, some of the underlying assumptions will break down. That said, I encourage you to try it out on a system where you bias 1 cv, and compare the reweighted FES as a function of this CV with another estimate (eg. eq 18 of Branduardi, Bussi, Parrinello JCTC 2012). I would be curious to see how it compares.

Pratyush

--

You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.

To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/e2969d31-c739-4328-8c33-f4fe4ea56cc1%40googlegroups.com.