aspirincode / one_step_retrosynth_ai

A novel, template-free approach for the single-step retrosynthesis task.

Single-step Retrosynthesis Prediction based on the Identification of Potential Disconnection Sites using Molecular Substructure Fingerprints

A novel, template-free approach for the one-step retrosynthesis task described in a journal paper which is currently under review. Depending on the journal decision, the link will be posted here.

Running the Code

This repository is still under construction, but it requires just a few more updates and it will be fully finished. The main functionalities are now available to be ran as scripts, which is described later in the document. There is still a high possibility of encountering bugs.

Installation

The necessary Python environment can be set-up using conda by simply running:

conda env create -f environment.yml

conda activate one_step_retrosynth_ai

If you encounter errors or conflicts for whatever reason, you can manually re-construct the environment. The following base libraries were used for the realization of the project:

• python: 3.6.10
• tensorflow-gpu: 1.12.0
• rdkit: 2020.03.3.0
• numpy: 1.16.0 (NOTE: This version is preferred to avoid annoying TF warnings.)
• pandas: 1.1.3

Additional libraries that are necessary for the code to be fully functional are:

• tqdm: 4.50.2
• scikit-learn: 0.23.2
• imbalanced-learn: 0.7.0
• matplotlib: 3.3.1
• cairosvg: 2.4.2

Everything else will be installed automatically as requirements for the base libraries.

Configuration

The general configuration of each step is stored in the config.json file. It specifies all of the necessary dataset, descriptor and model configurations.

1. Dataset Preparation

The dataset can be generated by running the following command:

python -m scripts.prepare_dataset config.json

The process consists out of 5 steps and the final dataset is saved in the folder specified in the configuration.

2. Model Training

The specified model can be trained and assessed by running the following command:

python -m scripts.train_model config.json

All of the hyper-parameters are specified in the configuration.

3. Full Pipeline Application

NOTE: Still under construction. Requires additional parameter updates and cosmetics. The full single-step retrosynthesis pipeline can be assessed by running the following command:

python -m scripts.run_evaluation config.json

All of the hyper-parameters are specified in the configuration.

Contact

For any questions and inquiries please feel free to contact the authors.