GitHub - a5hwinjs/Drug-Discovery-Using-GAN: So I kinda Explored the path of GAN and LSTM and built a Simple model for Discovering Drugs.

a5hwinjs / Drug-Discovery-Using-GAN Public

Notifications You must be signed in to change notification settings
Fork 1
Star 0

So I kinda Explored the path of GAN and LSTM and built a Simple model for Discovering Drugs.

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
.ipynb_checkpoints		.ipynb_checkpoints
Hack_GAN.ipynb		Hack_GAN.ipynb
LSTMs_create.ipynb		LSTMs_create.ipynb
Molecules Generation to Aid Drug Design.pdf		Molecules Generation to Aid Drug Design.pdf
README.txt		README.txt
dataset_orig.pdf		dataset_orig.pdf
drugs_sub.smi		drugs_sub.smi
fin1.csv		fin1.csv
hack_gyan.py		hack_gyan.py
lig_data.csv		lig_data.csv

Repository files navigation

The LSTMs_create.ipynb file implements our LSTM approach to the solution.
It works on the lig_data.csv(from the dataset_orig.pdf, from a paper by Stahl) dataset that contains only 335 data elements.


The Hack_GAN.ipynb showcases the usage of GANs in countering the same problem.
We tried it both on the lig_data.csv dataset and the fin1.csv(which we obtained from various drugs_smi.smi file, which we found on github)


Both can be executing independently and both provide the necessary outputs.