/aidd-workshops

This repository contains setup instructions and the notebooks for the AI in drug discovery workshops

Primary LanguageJupyter Notebook

Digital Chemistry: AI in Drug Discovery - Workshop

This repository contains the code and instructions for setting up for the AI in Drug Discovery workshop sessions.

Contents

During the workshop, we will be using a framework for machine learning in drug discovery called DeepChem. DeepChem is a Python library and toolchain that combines multiple machine learning algorithms, frameworks and datasets and provides a unified interface for training and evaluating models in drug discovery.

For each workshop session, we will be practically using DeepChem through a set of tutorial Jupyter Notebooks from the DeepChem library.

Workshop 1: Introduction to DeepChem (22/02/2023)

This workshop will introduce you to the basics of DeepChem and will allow you to familiarize yourself with the library.

  1. Basic tools of the deep life sciences

    Open In Colab

  2. Working with datasets

    Open In Colab

  3. An introduction to Moleculenet

    Open In Colab

  4. Molecular fingerprints

    Open In Colab

Workshops 2 and 3: Ligand and structure based predictive modelling (23/02/2023 and 01/03/2023)

This workshop covers lectures 3-5 of the AI in Drug Discovery lecture course.

  1. Creating a High Fidelity Dataset from Experimental Data

    Open In Colab

  2. Creating Models with TensorFlow and PyTorch

    Open In Colab

  3. Going Deeper On Molecular Featurizations

    Open In Colab

  4. Working With Splitters

    Open In Colab

  5. Introduction to Graph Convolutions

    Open In Colab

Workshop 4: Modeling protein-ligand interactions (03/03/2023)

  1. Modeling Protein-Ligand Interactions

    Open In Colab

  2. Modeling Protein-Ligand Interactions with Atomic Convolutions

    Open In Colab

Getting Started

In order to get started it is recommended you clone this repository locally, even if working in google colab. This will download the repository content to your local machine. To clone the repository locally, open a terminal (command prompt/powershell on windows) and type:

git clone https://github.com/GouldGroup/aidd-workshops.git

you should then have a new directory called aidd-worshops, navigate to this directory by typing:

cd aidd-workshops

Before each workshop session, this repository will be updated with the new workshop content. To update your local copy of the repository, open a terminal in the current directory (command prompt/powershell on windows) and type:

git pull

for each workshop session, you will find the relevant Jupyter Notebooks within the respective workshop directory in this repository (e.g. Workshop_1/The_Basic_Tools_of_the_Deep_Life_Sciences.ipynb). You will also find google colab links to the notebooks in the contents section of this readme.

If you choose to work in colab (recommended), you can open the notebooks in colab by clicking on the links in the contents section of this readme. You will need to be signed in to your google account. Should you wish to work locally, follow the instructions in the Setup section below.

Setup

To setup the development environment for these workshops, you need to have some conda distribution installed. We will use Miniconda for this workshop.

Install Miniconda

Download and install the relevant version of Miniconda for your operating system. (you are probably running a Windows 64-bit or MAC 64-bit system).

Install Dependencies

To install the required dependencies for the workshop and create a development environment for DeepChem, open a terminal in the current directory (Anaconda Prompt on Windows) and type:

conda env create -f environment.yml

Activate the Development Environment

you will then need to activate the newly created environment. To activate the environment type:

conda activate deepchem

you will now be able to run Jupyter notebook and open the notebooks in this repository!