DLAM: Essential Protein Predictions from Multi-Source Biological Data

A 366.0 KB Excel file containing results from the DLAM deep learning model for predicting essential proteins. The dataset was created by Shunxian Zhou and last updated on April 14, 2026. It presents model performance metrics from evaluations on the DIP and BioGRID protein interaction datasets.

Use Cases

• Benchmarking new essential protein prediction models based on reported ROC-AUC and AP scores.
• Training models that integrate multi-source biological information based on the described domain, localization, orthology, and expression features.
• Analyzing the impact of class imbalance on prediction performance based on the discussion of joint modeling.
• Comparing deep learning architectures for biological networks based on the comparison with methods like TCBB2021 and EPGAT.

Strengths

• Model performance is quantitatively reported with metrics like ROC-AUC, AP, and F1-score.
• Evaluation was conducted on two established protein interaction datasets: DIP and BioGRID.
• Uses a CC-BY-4.0 license, allowing for broad reuse and modification.

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 (366.0 KB), indicating limited scope, likely containing summary results rather than raw training data.

Provenance

Source

figshare, author Shunxian Zhou.

Collection Method

Results from a deep learning framework (DLAM) integrating domain composition, subcellular localization, orthology, gene expression, and a protein-protein interaction network.

Time Range

null

Freshness

Last updated 2026-04-14 22:01:28; freshness should be verified.

Geography

null