Compound-Protein Affinity Predictions for 81 Reproductive Toxicity Targets

Yu Ma's dataset contains predicted compound binding affinities for 81 reproductive toxicity targets, generated using a classification pretraining-regression fine-tuning framework. The data is stored in an XLSX file of 344.0 KB and was last updated on 2026-04-15. The framework demonstrated an average R² improvement of 0.324 under stringent testing and achieved a best model performance of R² = 0.797.

Use Cases

• Training machine learning models for target-based affinity prediction based on the described pretraining framework.
• Identifying structural alerts for reproductive toxicity based on predicted target affinity spectra.
• Investigating key protein targets for sex-specific reproductive diseases based on the predicted compound affinities.
• Assessing the generalization of affinity prediction models to novel chemical scaffolds using the provided evaluation results.

Strengths

• Includes predictions for a specific set of 81 reproductive toxicity targets.
• The generating model achieved a reported R² of 0.797 and MSE of 0.370 for the best-performing configuration.
• The framework demonstrated a substantial average R² improvement of 0.324 under the most stringent test scenario.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is small in scale at 344.0 KB, indicating limited scope.

Provenance

Source

figshare, author Yu Ma.

Collection Method

Predictions generated by a computational framework leveraging large-scale binary activity data for pretraining.

Time Range

null

Freshness

Last updated 2026-04-15 06:44:14; freshness should be verified.

Geography

null