ML-derived p-cymene derivatives using REINVENT4

An ML tutorial using the open-source REINVENT4 package with pretrained models

By Sofia Kiriakidi

In this tutorial, we will see how to use the open-source ML package REINVENT4, an AI molecular design tool for de novo design, scaffold hopping, R-group replacement, linker design and molecule optimization, developed by MolecularAI- Astra Zeneca, the original publication of which can be found here. We will combine it with Maize, a graph-based workflow manager for computational chemistry pipelines, developed by the same group and Autodock-GPU, developed by the Forli Lab at Scripps Research.

The most difficult part of this tutorial is to install all needed packages at your Linux machine and get them work together! Start by installing Maize, REINVENT4 and Autodock-GPU, following the instructions at their corersponding github pages, provided as links above. I am providing here a .yml for my working reinvent4 conda environment to use as a reference, but please have in mind that the aforementioned packages should be installed according to their instructions, you want be able to do it automatically using my .yml file (or if you can, email me how!!). The file can be found here.

Share: Twitter Facebook LinkedIn