Repository for the subject CDD

You will need rdkit (and potentionaly with postgres database), but Conda, Jupyter and Git are also very useful.

sudo apt install git

Download and install from git

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
# you will need probably to run:
# chmod u+x Miniconda3-latest-Linux-x86_64.sh to run this script
./Miniconda3-latest-Linux-x86_64.sh

Download and install miniconda

Download and install miniconda

git clone https://github.com/lich-uct/cdd2021.git
# create environment with rdkit, preferably updated version
# cdd is name of the environment
conda create -n cdd -c rdkit -c conda-forge rdkit=2019 jupyterlab 
conda activate cdd
jupyter lab # if you want to run it locally
jupyter lab --ip 0.0.0.0 --port XXXX # ip 0s to see jupyter on the Internet, port XXXX as maybe you don't want to run it on default 8888 port
# jupyter notebook --ip 0.0.0.0 --port XXXX alternative if you want only notebooks

On your part, please try to apply the workflows onto your particular set, and adapt them for your particular cases. Compare your results to the example case, and to those of your colleagues. We can discuss the results directly in your Jupyter notebooks. If you get stuck at any point, or want some additional information, please do not hesitate to ask - this course is for you.

The example jupyter notebooks with the workflows will be uploaded into the /exercises folder of the repository. I will upload the first notebook soon.

Make your own folders in the root of the repository; 'otavat' and 'rodionoa' already have one. Place your own data and notebooks and all other things you need in there, and upload them into this repository at your convenience. You already know git, so this should by no problem... git pull, git commit -m "my message", git push , the usual. Please keep the changes within your own directory, having to merge jupyter notebooks is a pain in the ass.

• exercises/01_getting_started: How to parse structures into RDKit, visualize them, run their methods

You will work on your projects, or 'zapoctovy projekt', projects will involve roughly the following:

• Picking a set of known biologically active structures, load and process it using RDkit
• Performing a basic analysis of the set, descriptors, scaffolds, fingerprints, similarities, etc.
• Comparing the properties of your picked set with other sets
• Trying out some standard cheminformatic methods on the sets
• Preparing set visualizations
• Making a basic classifier related to your picked and prepared dataset

The extent and exact form of all above points will be determined ad-hoc, depending on time, situation, progress, etc. We will periodically upload jupyter notebooks that describe the steps of the next task, the reasoning behind them, and their demonstration on an example set.

• Getting started with RDKit in Python
• RDKit documentation