Document
If you use IIDL-PepPI for research, please cite this paper:
Shutao Chen, Ke Yan, Xuelong Li, and Bin Liu*. Protein language pragmatic analysis and progressive transfer learning for profiling peptide-protein interactions. ( Submitted )
The dataset used in this study is based on the RCSB PDB collection and can be downloaded from the following links:
Randomly Split Datasets (5-fold):
5-fold Randomly Split Datasets.tar
Clustering Split Datasets (9-fold):
The source code of IIDL-PepPI can be downloaded from the following link:
Souece code of IIDL-PepPI:
The comparison file used by IIDL-PepPI to reduce the dimensionality of pre-training (ProtBERT) features can be downloaded from the following link (about 3G for protein, 200M for peptide):
File used by IIDL-PepPI for feature dimensionality reduction:
protein_webserver.pkl
peptide_webserver.pkl
The above file need to be downloaded and stored in the IIDL-PepPI/saved_models/protbert_feature_before_pca path for analysis.