Skip to content

kumar1202/Drug-Discovery-using-GANs

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Drug-Discovery-using-GANs

  1. We have implemented a novel computational strategy for design of drug molecules with the desired properties.The model implements a deep learning and reinforcement learning approaches.

  2. The model integrates two different deep neural networks - generator and discriminator with reinforcement learning.

  3. These two models are trained seperately but are used jointly to generate novel targeted chemical libraries.

  4. Our model uses string representation of molecules by their simplified molecule-input line entry-system(SMILES).

GAN_RL

  1. In this system, the generative model is used to produce novel chemically feasible molecules, that is, it plays a role of an agent, whereas the discriminative model (that predicts the properties of novel compounds) plays the role of a critic, which estimates the agent’s behavior by assigning a numerical reward (or penalty) to every generated molecule.

  2. The reward is a function of the numerical property generated by the predictive model, and the generative model is trained to maximize the expected reward.

  3. In the first phase of the method, generative and discriminator models are trained separately with a supervised learning algorithm.

GAN_RL

Yellow colored vioilin plot determines druglikeliness of the base dataset whereas the other determines the same for the generated data by GAN.

  1. In the second phase, both models are trained jointly with the RL approach to bias the generation of new chemical structures toward those with the desired physical and/or biological properties.

  2. In a nutshell we have used GAN model integrated with RL to design chemical molecules with specific range of physical properties, such as solubility, novelty, synthesizability.

About

Check out the dash visualization at https://dash-drug-explorer.plot.ly/out

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages