NIR-II Design

Prediction model used in the paper: Accelerated Design of Near-Infrared-II Molecular Fluorophores via First-Principle Understanding and Machine Learning.

Prediction Demo and Data Availability

• Access the jupyter notebook to view the demo on HOMO LUMO energy gap predictions with our trained model.
• The 24 NIR-II fluorophore cores and their predictions are available in the current Predictions.csv

Guide: Energy gap predictions

You can also use our trained model to make energy gap predictions given a NIR-II fluorophore SMILES input.

1. Download this repository and ensure all packages required are installed
2. Start the jupyter notebook and follow the steps (particularly, edit the To_Predict.csv file with the SMILES you need to predict, shown below)

3. The new predictions will then be saved in Predictions.csv

Requirements

• numpy
• pandas
• Tensorflow
• RDKIT
• Deepchem 2.4

Authors

Xu Shidang, Cai Peng Fei, Li Jiali,