biotech Deep-Interact Studio

Benchmark Datasets

Published benchmark datasets for each task, ready to download and train on.

Upload a file to the matching model builder and map its columns there. The label column is spelled lable in these files; 1 means interaction and 0 means no interaction.


PPI

Protein-Protein Interaction

Columns: ProteinA, ProteinB, lable

Human PPI Benchmark

Human protein pairs labeled as interacting or non-interacting.

download Download CSV
Pairs
65,836
Positive
32,224
Negative
33,612
File size
63.2 MB

File: PPI-processed.csv

Paper: Large-Scale Prediction of Human Protein-Protein Interactions from Amino Acid Sequence Based on Latent Topic Features

Pan, Zhang, and Shen, Journal of Proteome Research, 2010.


DTPI

Drug-Target Protein Interaction

Columns: SMILES, Protein, lable

NASNet-DTI Benchmark

Drug SMILES and protein sequences labeled for interaction prediction.

download Download CSV
Pairs
27,464
Positive
13,830
Negative
13,634
File size
16.0 MB

File: DTPI-processed.csv

Paper: NASNet-DTI: accurate drug-target interaction prediction using heterogeneous graphs and node adaptation

Zhong and Du, Briefings in Bioinformatics, 2025.


RPI

RNA-Protein Interaction

Columns: RNA, Protein, lable

RNA-Protein Benchmark

RNA and protein sequence pairs labeled as interacting or non-interacting.

download Download CSV
Pairs
20,824
Positive
10,412
Negative
10,412
File size
47.0 MB

File: RPI-processed.csv

Paper: RNA-protein interaction prediction using network-guided deep learning

Liu, Jian, Zeng, and Zhao, Communications Biology, 2025.