search

MDAnalysis / binder-notebook

Notebooks for interactive use on https://mybinder.org/
https://mybinder.org/v2/gh/MDAnalysis/binder-notebook/master?filepath=notebooks
GNU General Public License v3.0
4 stars 7 forks source link

data files for on-the-fly transformations missing #13

Closed orbeckst closed 5 years ago

orbeckst commented 5 years ago

The notebook transformations/on-the-fly-transformations.ipynb (open in nbviewer) (which was added in PR #10 ) uses 

u = mda.Universe('pept_in_memb.tpr', 'pept_in_memb.xtc')

These files pept_in_memb.tpr', 'pept_in_memb.xtc' are not installed with binder so the notebook will not run on binder. All notebooks should be fully self-contained.

Possible solutions:

orbeckst commented 5 years ago

@davidercruz can you please comment on this issue? How big are the files that you used for the example or are they already in MDAnalysisTests?

davidercruz commented 5 years ago

Hi @orbeckst, sorry for the late response. The files are really small - about 250KB each. Where do I upload them?

orbeckst commented 5 years ago

Can you please provide a very brief description of the trajectory? I am trying to decide if they are better suited for MDAnalysisData than the tests.

@jbarnoud your opinion is also welcome!

davidercruz commented 5 years ago

It's a 19 residue peptide embeded in a 128 DMPC membrane, without water molecules. The trajectory has 10 frames, 6085 atoms. It's a simple systems that I put together for the notebook, as I found it more interesting in the context of some of the on-the-fly transformations. It could be used in the tests, although I managed to do the tests with the existing test data.

orbeckst commented 5 years ago

Given that the tests don't require it, it sounds like a useful system for MDAnalysisData – small membrane system, stuff moving across boundaries, simple peptides and lipids. Would you be willing to share an extended version of the trajectory under a suitable open licence? Along the lines of

This would be a really good system to have available. We currently don't have anything with water + lipids + protein.

They way to make these available is to dump the files (TPR and XTC) on figshare. As an example see https://figshare.com/articles/Molecular_dynamics_trajectory_of_membrane_protein_NhaA/7185203/2 I will then add the data set to the available data sets or if you want to do the PR, I'll tell you what to do.

Once this is done, I'll add MDAnalysisData to the binder repo.

orbeckst commented 5 years ago

@davidercruz if creating the more extended trajectory is too much work then could you please nevertheless put your original trajectory on figshare? You are the author of the files so you have to do it and decide on a license. Thanks.

orbeckst commented 5 years ago

ping @davidercruz

orbeckst commented 5 years ago

@jbarnoud @mnmelo do you happen to have access to the file that @davidercruz has used here, or can you get in touch with him?

davidercruz commented 5 years ago

Hi, sorry for being MIA. Here are the files: https://figshare.com/articles/Short_molecular_dynamics_of_a_peptide_inside_a_pure_DMPC_membrane/8046437

orbeckst commented 5 years ago

Awesome! Many thanks.

I'll ~open~ update the issue for MDAData (https://github.com/MDAnalysis/MDAnalysisData/issues/34).

orbeckst commented 5 years ago

We now have a 0.7.0 release of MDAnalysisData that contains the new Membrane peptide dataset. Once the conda package has been built, we can add it to the binder setup file environment.yml and update the notebook.