DMAPLM: Benchmark Dataset for Drug-Disease Association Prediction

A benchmark dataset compiled for evaluating the DMAPLM multimodal pretrained framework for drug repositioning. The dataset was created by Hailin Chen and last updated on April 22, 2026. It is a small dataset of 5.5 KB, stored in an XLS file format.

Use Cases

• Benchmarking new drug-disease association models based on the described evaluation framework.
• Training machine learning classifiers for drug repositioning based on multimodal features.
• Conducting case studies on novel drug-disease relationships based on the dataset's predictions.
• Analyzing model performance in cold-start scenarios based on the described experimental setup.

Strengths

• Dataset is licensed under CC-BY-4.0, allowing for open sharing and adaptation.
• Serves as a benchmark for a model achieving an AUROC of 0.8919 and AUPR of 0.9116.
• Associated with a framework demonstrating robust performance in cold-start scenarios.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The 5.5 KB size suggests a very limited scope, likely containing aggregated metrics or a small sample.

Provenance

Source

figshare, author Hailin Chen.

Collection Method

Compiled as a benchmark for evaluating the DMAPLM framework, likely derived from public biomedical sources.

Freshness

Last updated 2026-04-22 17:40:50; freshness should be verified.