Pre-trained model weights and data of TamGen

We are pleased to release supplementary materials for the paper `TamGen: Drug Design with Target-aware Molecule Generation through a Chemical Language Model`.

1. TamGen-code.zip: The code for TamGen. For the latest version and updates, please refer to `https://github.com/SigmaGenX/TamGen`. 
2. crossdocked_results.zip: Decoding results from TamGen and related baseline models. This file also includes the code to reproduce Fig. 2(b).
3. crossdock_case_study.Fig2d.txt: It contains the SMILES strings from the case study presented in Figure 2(d).”
4. SourceDataFile.xlsx: Source data referenced in the paper.
5. DockingPose.zip: Docking poses corresponding to Figure 5 in the paper.
6. gpt_model.zip: The pre-trained GPT-style model.
7. checkpoints.zip: Pre-trained checkpoints for model inference.
8. pdb_ids.csv: includes pdb_ids of additional PDB files used for training the model, which are utilized for designing compounds targeting tuberculosis (TB). 
9. compounds_designStage.txt: 2612 compounds generated in the `Design` stage of Figure 3.
10. compounds_refineStage.txt: 8365 compounds generated in the `Refine` stage of Figure 3.