End-to-end ML pipeline for drug discovery from generating novel small molecules with desired effects to predicting the synthesis (retrosynthesis) pathways of those generated molecules.
- Website
- Pitch Video (4:40)
- Demo Video (1:03)
- Article about Reinforcement Learning
- Article about VAE model
- Older Version of Synbiolic's Open Source Repo at William Law's Github
- Microsoft Imagine Cup (IC) N.A. Regional Finalist (Top 5 of 28000)
- Microsoft IC Feature
- Runner Up Microsoft IC Feature
- Microsoft IC Pitch
git clone https://github.com/joeym-09/Synbiolic.git
cd synbiolic
npm install
cd client
npm install
cd ../
npm run dev
- The number of recommended molecules to generate is 20-50 (for demo purposes).
- It will then take you to all the generated molecules - where you can view the molecules and the plc-50 distribution
- You can save any of the molecules and click on them to bring up the retrosynthesis pathway
- The retrosynthesis pathway will have more details about the molecule itself