Datasets from "A generative deep learning approach to de novo antibiotic design"

This repository accompanies the manuscript "A generative deep learning approach to de novo antibiotic design" by Krishnan, Anahtar, Valeri, et al., Cell 2025 , and contains: 

• Thee checkpoints for all Chemprop models used in this study:
  - Neisseria gonorrhoeae (NG)
  - Staphylococcus aureus (SA)
  - Hepatocellular carcinoma (HepG2)
  - Human skeletal muscle cells (HSkMCs), and
  - Human lung fibroblasts (IMR-90)
• List of generated molecules in a fragment-based way (with fragment seeds: F1, F2 and F2') using F-CReM and F-VAE as well as those generated entirely de novo with CReM and JT-VAe are given. See the main text for further details
• PRIDE access to the Raw Proteomics Data
• Raw Transcriptomics Data
• Synthesis routes, NMR and mass spectra of the lead compounds and their analogs

All code used in this manuscript can be found at https://github.com/aartikrish/de-novo-antibiotics.