This repository contains executable notebooks and links related to the OSM Series 4 work by Ersilia.
The current manuscript is publicly visible in this online document. We expect to submit it as a Note to ACS Medicinal Chemistry Letters in response to this call for papers related to AI.
Author guidelines can be found here and a short comment on how to increase chances of acceptance can be found here.
In sum:
- Abstract: 150 words.
- Text: 2500 words, no sections.
- References: 30.
- Figures and tables: 2-4.
- Figures format: One column: up to 240 pt (3.33 in). Two columns: between 300 and 504 pt (4.167 in. and 7 in.). Lettering no smaller than 4.5 pt.
- OSM Issue #34
- Generative modelling, round 1
- Generative modelling, round 2
- Predictive model
- Selection of candidates for synthesis
- PfATP4 AlphaFold Structure
- ChemSampler
Create a conda environment and activate it.
conda create -n osms4 python=3.11
conda activate osms4
Install Ersilia Compound Embeddings and after you have installed it.
cd ~/Desktop
git clone https://github.com/ersilia-os/compound-embedding.git
cd compound-embedding/lite
pip install .
cd ../..
rm -r compound-embedding
Install other necessary Python libraries.
pip install pandas
pip install umap-learn
pip install statsmodels
pip install git+https://github.com/ersilia-os/stylia.git
In the notebooks
folder, there are two Juypter Notebooks that are used to produce the display items of the manuscript.
ChemicalSpaceExploration.ipynb
is used for the main Figure 1.CompetitionBenchmark.ipynb
is used for the supplementary Figure S1.