Data and Code for :Constrained SELFIES-Based Inverse Design of Ionic Liquids for Fuel Denitrogenation
Description
This dataset contains the input workbooks and Jupyter notebooks required to reproduce the constrained and unconstrained SELFIES-based ionic-liquid generation and ranking workflows. The files density_P.xlsx and viscosity_P.xlsx provide the literature-derived ionic-liquid ion records used to construct the parent/seed ion pool for SELFIES mutation. The file All_contaminants_decane_2_with_ion_smiles_CHARGE_OK.xlsx contains the charge-checked cation/anion SMILES and COSMOtherm/COSMO-RS thermodynamic data used for the decane, quinoline, and carbazole ranking tasks. This workbook includes the quinoline and carbazole sheets, which are filtered by the notebooks during model training and deployment. The notebook Constrained mutations and relevant codes.ipynb implements the constrained workflow, including ion standardization, family selection, SMARTS/MCS family-core extraction, SELFIES token mutation, charge and core-retention checks, descriptor-envelope filtering, post hoc cation cleanup, PubChem CID support filtering, constrained cation–anion pair construction, two-stage learning-to-rank, and deployment of the trained ranking models. The notebook Unconstrained mutation ablation and related codes.ipynb implements the relaxed comparison workflow used to evaluate the effect of removing the stricter chemical constraints. This workflow retains the core family and charge-localization rules while omitting the descriptor-envelope, post hoc cleanup, and PubChem-support filters used in the constrained manifest. The notebooks were run in Google Colab using an NVIDIA A100 GPU runtime for the ranking-model sections. File paths may need to be updated to match the user’s working directory; keeping the uploaded filenames unchanged is recommended. The uploaded workbooks allow reproduction of the generation, filtering, ranking, and deployment workflow. Independent re-execution of COSMO-RS/COSMOtherm calculations requires access to licensed BIOVIA COSMOtherm software.
Files
Institutions
- Khalifa University of Science and TechnologyAbu Dhabi, Abu Dhabi