Blog

MDAnalysis + Machine Learning Workshops

In collaboration with CCPBioSim, 3 MDAnalysis core devs (@ialibay, @micaela-matta, @richardjgowers) together with @ppxasjsm (University of Edinburgh) and @degiacom (University of Durham) organised and ran a 2-day hybrid (in person/online) workshop titled “MDAnalysis and Machine Learning for Molecular Simulations” on June 9-10 2022, immediately following the 8th Annual CCPBioSim Conference. The workshop received generous funding from the MGMS Early Career Workshop Initiative.

The topics covered included:

  • The fundamentals and basics of MDAnalysis
  • How to get started with machine learning and tools such as scikit-learn
  • How to use MDAnalysis in conjunction with machine learning
  • Applying these tools to your own research projects

The workshop was hosted at the University of Edinburgh and consisted of 4 half days in a hybrid format, with over 40 in person attendees and over 30 attendees online at any point in time. A special thank you to Jasmin Güven (PhD student in the group of Antonia Mey) who moderated questions from online participants!

The teaching materials can be run on Google Colab and are found on GitHub.

The workshop style was massively adapted from Software carpentries, using a mixture of life coding, walking through Jupyter notebooks and problem sections. We used the post-it system for instant feedback from participants, and also ran a post-workshop survey, collecting overwhelmingly good feedback.

Thanks to this success, @micaela-matta, @degiacom and @ppxasjsm will be running the workshop again at the CCP5 Summer School on July 26-27 2022, in Durham!

We are glad to see such a high demand for training resources and events from the molecular simulation community. We look forward to continuing to deliver these workshops in the future.

workshop instructors introducing MDAnalysis

— the MDAnalysis team

SciPy 2022

MDAnalysis recently attended SciPy 2022!

SciPy was a fantastic opportunity to engage with the open-source community, build connections and see the amazing work being conducted all across the Scientific Python ecosystem.

Additionally, Hugo MacDermott-Opeskin presented a poster on our ongoing work as part of our CZI-EOSS4 grant. You can find his poster at the following FigShare DOI.

As part of the SciPy sprints session we ran an MDAnalysis Sprint where we invited people to contribute to MDAnalysis and collaborated on various ideas with community members. We would like to thank members of the Zarr, Xarray, Dask, scikit-learn and Pangeo Forge communities for some helpful and stimulating discussions.

We also participated in a #bio_at_scipy meetup to share ideas around engaging with the broader community and encouraging contributions.

— Richard Gowers, Hugo MacDermott-Opeskin, Tyler Reddy

Apple M1 conda packages for MDAnalysis 2.2.0

We now also have conda-forge packages for our MDAnalysis 2.2.0 release that directly support the Apple M1 ARM architecture (labelled osx-arm64).

On all

  • supported Python versions (3.8, 3.9, 3.10)
  • supported Operating Systems (Linux, Windows, MacOS)

you are now able to install and upgrade with conda

conda update -c conda-forge mdanalysis

For everything else about the new release, read our blog post about MDAnalysis 2.2.

— The MDAnalysis Team