becksteinlab / adkgromacstutorial Goto Github PK

see http://www.sphinx-doc.org/en/stable/markup/code.html

We can use travis to generate the docs and push to the gh-pages branch (or use master directly) or use Read The Docs.

In any case, the docs should be live somewhere.

As noted on the gmx-users list, the tutorial does not work with Gromacs 2018:

Missing index groups

On 7/3/18 8:32 AM, Mahmood wrote:
The tutorial contents look a bit outdated
Well the main page says that the release date is Jul 3 and it is compatible with 2018.I changed tau-p to just one value (tau-p = 2.5)

Well, the home page says the tutorials "should be compatible" with the listed versions, but there are several things that I note that won't work with 2018, including a few commands. I encourage you to provide feedback to the author(s) to note the problems with 2018, as surely you have had to change a few commands, e.g. to include -r (which is now mandatory for running grompp with restraints).

Now, I get
Program: gmx grompp, version 2018
Source file: src/gromacs/gmxpreprocess/readir.cpp (line 2682)

Fatal error:
Group PROT referenced in the .mdp file was not found in the index file.
Group names must match either [moleculetype] names or custom index group
names, in which case you must supply an index file to the '-n' option
of grompp.

Again it seems there are inconsistencies in the input files. I only took a quick look through the tutorial, but I see no mention of making any of the index groups used in comm_grps in md.mdp.

MDP options

On 7/3/18 7:48 AM, Mahmood wrote:
Hi,Following adkgromacstutorial, at the energy minimization step [1], I hit an error. All previous steps are fine.

Command line:
gmx grompp -f md.mdp -p ../top/4ake.top -c ../posres/posres.pdb -o md.tpr -maxwarn 3

Yikes! -maxwarn 3? Very dangerous... :)

ERROR 1 [file md.mdp, line 94]:
Right hand side '2.5 2.5' for parameter 'tau-p' in parameter file is
not a real value

Although it says line 94, I see that parameter is in line 76
$ grep -n tau-p md.mdp
76:tau-p = 2.5 2.5 ; Barostat coupling constant (ps) – Piggot (5.0)

Any idea about that?

[1] https://adkgromacstutorial.readthedocs.io/en/latest/simulation.html

The tutorial contents look a bit outdated (Oliver did just warn about this in another message). tau_p only takes one value. In older versions, additional values were silently ignored, but not in recent GROMACS versions.

If you're trying to use version 2018, may I suggest http://www.mdtutorials.com/gmx/index.html

-Justin

push local changes!

Overview

The initial version of the tutorial that is live on RTD as of March 28, 2017 is a basic version that could be followed by anyone, in principle. The purpose of the advanced tutorial would be to make it as simple as possible for lab members to get familiar with Gromacs, with the useful side-effect of generating data that could be used for publication. To create the advanced AdK Gromacs tutorial, I intend to piece together the scripts I've used (also Taylor) for AdK simulations. This advanced version should:

• Do proper multi-step minimization/equilibration
• Include the bash scripts that help automate the process—from system setup to production runs—to ensure consistency of results
• Have details and information that would be specific to (new) members of Becksteinlab (e.g., module load gromacs/5.1.3/...)
• Other things I might have forgotten

Separate repository?

I'm not sure if we should have a separate repository for this version. On the one hand, it may be easier to maintain separate repos, make it easier to deal with Read the Docs by keeping each tutorial cleaner (fewer toctree levels), and allow the current basic tutorial to stand on its own so that others outside the lab can use it. On the other hand, it seems a bit excessive to have, say, "AdKGromacsTutorialBasic" and "AdKGromacsTutorialAdvanced"; also, Gromacs tutorials already exist, so we'd only really be offering an alternative. Any thoughts on the matter?

The tutorial files that are included in the repository need to be organized for consistency with the documentation (now on RTD). The coord, emin, MD, posres, solvation, and bin directories also do not really need to be there, since the user is expected to generate this directory structure on their own. Also, only the *.mdp files that are referenced in the basic tutorial docs should be included in the repository—some unnecessary ones are lying around.