SAR-DRG: 96,158 Affinity-Labeled R-Chain Samples for Lead Optimization

96,158 samples across 29,387 scaffold-pocket groups provide a foundation for structure-activity relationship (SAR) analysis. This dataset, authored by ChaseCheng and hosted on Hugging Face, is designed for realistic R-chain generation in lead optimization. It was last updated on May 9, 2026.

Use Cases

• Training generative models for R-chain enumeration based on shared scaffold-pocket contexts.
• Evaluating molecular generation algorithms using affinity-labeled samples.
• Analyzing structure-activity relationship (SAR) patterns within defined chemical groups.

Strengths

• Contains 96,158 total samples.
• Organized into 29,387 distinct scaffold-pocket groups, providing structured context.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Hugging Face dataset authored by ChaseCheng.

Collection Method

Likely compiled from chemical databases or computational simulations for scaffold-pocket analysis.

Freshness

Last updated 2026-05-09 11:10:08; freshness should be verified.