Day 3
Plans for Day 3
Today, we will begin by wrapping-up our agenda for Day 2. Specifically, we will finish submitting our equilibration runs, submit our first set of solvation free energy calculations, review some "theory" concepts from Day 1, and then review the content of our input files. Then what we will add that is new is computing the solvation free energy from our Gromacs trajectories using MBAR (multi-state Bennett's Acceptance Ratio method).
A copy of my presentation from today is provided below. For Day 3 there will not be a screen cast. The "theory" slides will be covered in the screen cast for my Day 1 presentation.
Setting Up Software to Perform Solvation Free Energy Calculations with MBAR
Note that this is not necessary for this workshop. I add this material here for your future reference. To perform our MBAR calculations, we will use the freely available software PyMBAR, a Python implementation of MBAR. Then to use MBAR to compute solvation free energies, we will use a set of programs distributed as alchemical-analysis. I encountered several challenges to get this up and running on OSC. First, alchemical-analysis is no longer under development, and you are instead referenced to a new, more general project alchemlyb. But at present, it is not clear how to perform the desired analysis. To use alchemical-analysis, they say to use PyMBAR version 3.0.0.dev0 or above. But when I used the latest version, 3.0.4, it did not work. I ended up using an older version 2.0.1 which worked! Copies of the PyMBAR and alchemical-analysis that I was able to get to work together are provided in the Google Drive folder below, and I additionally provide a screen cast of how I set-up the software on OSC. I expect this should be similar to your local computing center. Note that there are other approaches to set-up the software that you can find on their respective websites. However, with OSC I am restricted in my ability to install, which is likely the case at your shared computing center. You could alternatively download all of your Gromacs files and perform the calculations locally, which would give you some freedom with the install.
In addition to the websites for PyMBAR and alchemical-analysis, there are a few additional references that should be sited when using these tools which I provide below.
Shirts, M.R. & Chodera, J.D. J. Chem. Phys. (2008) 129: 124105. https://aip.scitation.org/doi/10.1063/1.2978177
Chodera, J.D., Swope, W.C., Pitera, J.W., Seok, C. & Dill, K.A. J. Chem. Theory Comput. (2007) 3: 26-41. https://pubs.acs.org/doi/10.1021/ct0502864
Klimovich, P.V., Shirts, M.R. & Mobley, D.L. J. Comput.-Aided Mol. Des. (2015) 29: 397-411. https://doi.org/10.1007/s10822-015-9840-9
Perform Solvation Free Energy Calculations with MBAR
Once your Gromacs free energy calculation simulations are complete, we can compute the solvation free energy using MBAR. A screen cast of exactly how to do so is provided below. You will notice that PyMBAR computes the solvation free energy using a number of methods. You are highly encouraged to use the MBAR value. Also, a feature of PyMBAR that I very much appreciate is its analysis of the correlation time of you system, and the use of only uncorrelated (independent) samples. This information will be printed to screen. Take note of this. For example, when I perform solvation free energy calculations in ionic liquids, they have much slower dynamics. Much longer simulation times are necessary to obtain sufficient independent samples for analysis. You are encourage to explore the effect of simulation time and number of states on the computed free energy.
As a final note, to facilitate performing calculations with PyMBAR, I use a script which properly adds the necessary software to my Python path (which is necessary if you built the software like me). A copy has been added to the Day 1 page, in the "lidocaine_example" folder on Google Drive. A copy is also in our shared scratch space in the "lidocaine_example" folder. Normally I would have put a copy in the "fe_calculation" folder for each system. However, at the time when I was preparing the example directory for the first day of the workshop, I was struggling to get the analysis software to work on OSC. It took me a great deal of time to get it working properly.
Summary of Status After Day 3
We had a great day today, but unfortunately again had internet issues. We were able to cover all of the material we had planned, except I was not able to demonstrate the calculation of solvation free energy using MBAR. This is shown in the screen cast posted above, and hopefully we can quickly demonstrate tomorrow.