GLASS 2.0

Download GLASS 2.0

Access comprehensive datasets from GLASS 2.0 for research and analysis. All datasets are freely available for academic and non-commercial use.

Raw Datasets

Complete original data from GLASS 2.0 database

Complete Dataset

glass2_full.tsv

~230 MB
Complete protein-ligand pair records with all experimental data, binding affinities, and metadata from GLASS 2.0 database.
1,147,227 entries Updated: 2025-06-27

Protein Information

protein.json

~3.8 MB
Comprehensive protein information including sequences, structures, gene names, organisms, and functional annotations.
3,308 proteins Updated: 2025-06-27

Ligand Basic Information

ligands.tsv

~210 MB
Basic molecular information including UCI, SMILES, InChI, names, molecular properties, and drug-likeness metrics.
458,292 ligands Updated: 2025-06-27

Cross References

uci_xref.json

~93 MB
Mapping from UCI identifiers to external database IDs including ChEMBL, PubChem, DrugBank, and other chemical databases.
458,292 mappings Updated: 2025-06-27

ML-Ready Datasets

Curated high-quality datasets optimized for machine learning applications

Classification Dataset

glass2_cls.csv

~60 MB
Binary classification dataset with manually curated positive/negative labels. Negative samples from experimental filtering and ChEMBL annotations.
Balanced dataset High quality

Active Compounds

glass2_reg_act.csv

~49 MB
Regression dataset containing active compounds with IC50/EC50/Ki/Kd values < 10,000 nM. Suitable for potency prediction models.
< 10μM cutoff Regression ready

Inactive Compounds

glass2_reg_inact.csv

~9.25 MB
Regression dataset containing inactive compounds with IC50/EC50/Ki/Kd values ≥ 10,000 nM. Complements active dataset for comprehensive modeling.
≥ 10μM cutoff Regression ready