abhik1368 / jaeger

JAEGER is a deep generative approach for small-molecule design

JAEGER

JT-VAE Generative Modeling (JAEGER) is a deep generative approach for small-molecule design. JAEGER is based on the Junction-Tree Variational Auto-Encoder (JT-VAE) method [1], which ensures chemical validity for the generated molecules.

JAEGER is trained on existing molecules associated with activity values measured in a given assay. During training, JAEGER learns how to map each molecule onto a (high-dimensional) coordinate space, often referred to as the latent space. JAEGER also learns how to decode a coordinate position in the latent space back to a molecule.

To generate new molecules, JAEGER defines numerical search strategies to efficiently and effectively explore that latent space. JAEGER couples the exploration of that latent space together with activity predictive models to discover and optimize novel active molecules.

[1] Wengong Jin, Regina Barzilay, Tommi S. Jaakkola: Junction Tree Variational Autoencoder for Molecular Graph Generation. ICML 2018: 2328-2337

Contents

• src: Python source code for JAEGER
• models: Data for building demo model

System requirements

Hardware

GPUs

We have tested JAEGER on machines with the following GPUs:

• NVIDIA Tesla K80
• NVIDIA Tesla V100

Software

Operating systems

We have tested JAEGER on machines with the following systems:

• Red Hat Enterprise Linux 6
• CentOS Linux 7

Sofware dependencies

• python 3.8.6
• pandas 1.1.5
• numpy 1.19.5
• pyjanitor 0.20.10
• pytorch 1.7.0
• rdkit 2020.09.3
• scikit-learn 0.24.0
• streamlit 0.74.1

Installation

• Install the python libraries mentioned in Software dependencies above into your python environment.

• Get the jaeger branch from the JT-VAE JAEGER fork located here. Include the icml18-jtnn and the icml18-jtnn/jtnn directories in your python path.

• Get a copy of the JAEGER repo (this repo). Include the src directory in your python path

Installation time of the software dependencies will vary depending on your computational environment. The larger packages like pytorch can take a couple of hours.

Demo dataset

We include a demo dataset with all molecules with measured 3D7 inhibition activity from the Deposited Set 2: Novartis GNF Whole Cell Dataset hosted at the ChEMBL - Neglected Tropical Disease archive. The demo dataset is located at ./models/training_data/Novartis_GNF.csv.

Training a model

See TRAINING.md

Generating molecules

See GENCHEM.md

License

Copyright 2021 Novartis Institutes for BioMedical Research Inc.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

See LICENSE.txt

Contact

[email protected]

Citation

Please cite the code as follows:

Godinez, W. J. & Ma, E. J. Novartis/JAEGER: Public. Zenodo, doi:https://doi.org/10.5281/zenodo.5794429 (2021).

DOI