NetCal-DTI: Benchmark Data for Drug-Target Interaction Prediction

Fanbo Ju's dataset aggregates multiple sources for benchmarking a network-calibrated drug-target interaction prediction model. It includes data from DrugBank V6.0, BIOSNAP, a BindingDB subset with Kd values, and single-cell transcriptomic data from GSE301188 for Tucidinostat-treated samples. The collection also contains supplementary results from ablation studies, similarity heatmaps, prediction scores, and docking simulations.

Use Cases

• Benchmarking inductive drug-target interaction models based on aggregated DrugBank, BIOSNAP, and BindingDB data.
• Analyzing the effect of network topology calibration using provided ablation study results.
• Screening potential drug targets using the Tucidinostat single-cell transcriptomic response data.
• Validating computational predictions with provided docking results for proteins CDK2 and SERT.

Strengths

• Integrates data from multiple established biological databases (DrugBank, BIOSNAP, BindingDB) and a public GEO dataset.
• Includes supplementary experimental results like ablation studies and docking simulations, adding validation context.
• Released under a permissive CC-BY-4.0 license, facilitating reuse.

Limitations

• Column names and specific row counts are not provided, limiting immediate understanding of data structure.
• The dataset appears to be a collection of results and source references rather than a single, cleanly structured table.
• The last update date (2026-03-21) is in the future, indicating a potential metadata error.

Provenance

Source

DrugBank V6.0, BIOSNAP, BindingDB subset, and NCBI GEO dataset GSE301188.

Collection Method

Aggregated from public repositories for benchmark testing.

Time Range

null

Freshness

Last updated 2026-03-21 13:04:30 (note: date is in the future).

Geography

null