ZINC15 for Drug Similarity Search
<p><strong>Abstract:</strong> This dataset is a subset of the ZINC15 database, specifically filtered and processed for molecular similarity search applications using MegaMolBART embeddings. The subset focuses on drug-like molecules with specific physicochemical and purchasability properties.</p> <p><strong>Keywords:</strong> ZINC15, Molecular Similarity Search, MegaMolBART, Drug Discovery, Cheminformatics.</p> <p><strong>Background:</strong> The ZINC15 database is a comprehensive collection of commercially available compounds for virtual screening. This subset was created to facilitate the development of machine learning models for drug discovery, particularly those based on molecular embeddings.</p> <p><strong>Methodology:</strong> The ZINC15 database was queried using the following criteria:</p> <ul> <li>Molecular weight <= 500 Daltons</li> <li>LogP <= 5</li> <li>Reactivity level = "reactive"</li> <li>Purchasability = "annotated" </li> </ul> <p>The resulting dataset was then processed to extract MegaMolBART embeddings for each molecule.</p> <p><strong>Data Description:</strong> </p> <p>The dataset is organized into three folders:</p> <ul> <li><code><strong>/data/project/ubrite/drg-depot/zinc15-similarity-search/raw-data/</strong></code> (66 GB): This folder contains the raw data files obtained from the ZINC15 database after applying the filtering criteria.</li> <li><code><strong>/data/project/ubrite/drg-depot/zinc15-similarity-search/processed-data/</strong></code> (13 GB): This folder contains the processed data, including the extracted MegaMolBART embeddings for each molecule. </li> <li><code><strong>/data/project/ubrite/drg-depot/zinc15-similarity-search/query/</strong></code>: This folder contains sample SMILES strings and their corresponding embeddings for performing similarity searches.</li> </ul> <p><strong>Technical Specifications:</strong> </p> <ul> <li><strong>Format:</strong> SMILES strings, numerical data (embeddings)</li> <li><strong>Size:</strong> 79 GB (total)</li> <li><strong>License:</strong> This dataset is derived from the ZINC15 database and processed using MegaMolBART. It is subject to the licenses of both the ZINC15 database and the MegaMolBART model. <ul> <li>ZINC15 Database: ZINC15 data is made available under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license. For more information, please visit the ZINC15 website.</li> <li>MegaMolBART: The MegaMolBART model and its associated data are copyrighted by AstraZeneca and NVIDIA. The usage of MegaMolBART is subject to the terms and conditions specified by the copyright holders.</li> </ul> </li> </ul> <p>By using this dataset, you agree to comply with the licenses and conditions imposed by the ZINC15 database and MegaMolBART.</p> <p><strong>Access and Usage:</strong> </p> <p>The dataset is available for download through Zenodo. Users are encouraged to acknowledge this dataset and the corresponding Zenodo entry in any publications or research projects that utilize the data. </p> <p><strong>Contact:</strong> Fuad Al Abir, fuad021@uab.edu</p>
ShareScore
32/100
Overall dataset sharing score
Score breakdown
These five areas show where the dataset supports — or may limit — practical reuse.
- Stewardship
- 4
- Harmonization
- 4
- Access
- 16
- Reuse readiness
- 8
- Engagement
- 0