Molecular dynamics (MD) simulation is one of the useful approach for estimating protein-ligand interaction. However, computational cost of MD simulation is high, so large computational resource is required when we run the large scale MD simulation.
On the other hand, now we can use freely available could computational resource such as Google colab. It’s not only limited CPU usage but also GPU. And recently there is really useful open source package for MD named OpenMM. Compared to commercial MD packages, openmm is required lots of steps for running MD. So it will be a hurdle for beginners like me ;)
Recently Pablo R. Arantes’s group published really useful article about running MD on cloud environment. The URL of publication is below.
And fortunately, the author shared code on github. It is really cool stuff isn’t it!
You can find source code is here. https://github.com/pablo-arantes/Making-it-rain
The repo provides example jupyter noetbook. So user can run MD simulation conveniently.
From the article, the speed of MD simulation is 230ns/day with P100GPU(Colab Pro version). The performance is different to free version of Colab but it useful for learning how to run MD with OpenMM.
I recommend to read original article to understand the details of the system. And I would like to say thank again the authors.