Using MDAnalysis for Efficient Simulation Pre- and Post-Processing

Contents

This repository contains the materials for the 3-day MDAnalysis 2021 PRACE/SURF workshop.

The workshop is structured in the following manner:

Day 1:

1. Day1-Session1-Lecture

• Welcome.ipynb

  Workshop welcome presentation.

• Lecture1_Molecules.ipynb

  The first lecture covering the general structure, objects and workflows of MDAnalysis.

2. Day1-Session1-Practical

• session1_practical.ipynb

  Practical complementing Day1-Session1-Lecture/Lecture1_Molecules.ipynb introducing the fundamental objects of MDAnalysis (e.g. Universe, AtomGroup, Atom, Residues, Segments).

3. Day1-Session2-Lecture

• session2-dynamics.ipynb

  Lecture covering the main concepts of coordinate I/O in MDAnalysis.

4. Day1-Session2-Practical

• session2.ipynb

  Practical complementing Day1-Session2-Lecture/session2-dynamics.ipynb. This practical demonstrates the main objects linked to coordinate I/O, and demonstrates how they can be used for basic trajectory analysis.

Day 2:

1. Day2-Session1-Lecture

• lib.distances.pdf

  Lecture covering both the MDAnalysis.lib.distances module and the MDAnalysis' analysis framework (e.g. AnalysisBase).

2. Day2-Session1-Practical

• day2session1-lib.distances.ipynb

  Practical complementing Day2-Session1-Lecture/lib.distances.pdf. Demonstrates how to calculate hydrogen bonds using both MDAnalysis.lib.distances and MDAnalysis.analysis.hydrogenbonds.

3. Day2-Session2-Lecture

• universe_creation_and_manipulation-lecture.ipynb

  Lecture covering how to;

  • Create, copy, and merge Universe objects
  • Manipulating positions (e.g. translate, rotate, PBC wrap/unwrap)
  • Talk to other libraries through MDAnalysis.converters

4. Day2-Session2-Practical

• day2session2-Analysis.ipynb

  Follow-up to Day2-Session1-Lecture/lib.distances.pdf and Day2-Session1-Practical/day2session1-lib.distances.ipynb. This practical demonstrates how one could write their own AnalysisBase-derived analysis class.

Day 3:

1. Day3-Session1-Lecture

• MDAParallelization.pdf

  Lecture convering various strategies for parallelizing MDAnalysis-based workflows.

2. Day3-Session1-Practical

• parallelism.ipynb

  Practical complementing Day3-Session1-Lecture/MDAParallelization.pdf. Here we demonstrate how both multiprocessing and MPI-based parallelism can be employed to analyze lipid headgroup orientation. Note: the MPI portion of this practical assumes access to SURF's LISA HPC platform, see Day3-Session1-Practical/README.md for more information.

3. Day3-Session2-Lecture

• more-features.ipynb

  Final lecture demonstrating some of the other features available in MDAnalysis such as auxiliary data reading, and on-the-fly transformations.

4. Day3-Session2-Practical

• creating_and_manipulating_universes.ipynb

  Practical complementing Day2-Session2-Lecture/universe_creation_and_manipulation-lecture.ipynb This practical demonstrates how to seemlessly interface RDKit and OpenMM with MDANalysis.

5. Day3-Closing-Lecture

• MDAnalysis_Workshop_final_no_logo.pdf

  SURF / PRACE workshop closing remarks.

Setting up your python environment

Instructions for setting up your environment to run this workshop locally are provided in INSTALL.md.

A full list of the required Python packages can be seen inside environment.yml.

Note: this workshop uses the beta release of MDAnalysis 2.0.0.

Course pre-requisites

The course assumes that attendees have a working proficiency in using Jupyter notebooks, Python (especially the NumPy library), and the bash shell.

Binder

The tutorial materials can be accessed online via binder. To launch the binder instance, click here.

License

The code examples in this repository are licensed under the GPL version 2.0 license. Non-code content is licensed under the Creative Commons Attribution-ShareAlike 4.0 International license. The MDAnalysis logo and its derivatives are licensed under the Creative Commons Attribution-NoDerivs 3.0 Unported License.

Acknowledgements

Please see AUTHORS.md for a list of contributors to the workshop materials.

Special acknowlegements go to the 2018 MDAnalysis Workshop and Hackathon which these materials build on.