High-Throughput Screening Assay Datasets from the PubChem Database.

Area of Science:

• * Cheminformatics
• * Computational Chemistry
• * Drug Discovery

Background:

• * Publicly available high-throughput screening (HTS) data is crucial for developing ligand-based computer-aided drug discovery (LB-CADD) methods.
• * Existing HTS repositories often present data in a convoluted manner, requiring extensive curation to identify reliable active and inactive compounds.
• * False positives and cross-reactivity in primary screens necessitate rigorous validation through secondary and confirmatory assays.

Purpose of the Study:

• * To illustrate a robust curation process for hierarchically related confirmatory screening data.
• * To describe a procedure for re-uploading curated findings into public repositories like PubChem.
• * To provide high-quality, publicly accessible datasets for benchmarking LB-CADD methods.

Main Methods:

• * Development and application of a data curation protocol for analyzing hierarchically related confirmatory screening assays.
• * Case studies involving two specific protein targets to demonstrate the curation process.
• * Re-uploading curated datasets to PubChem for public accessibility.

Main Results:

• * Successful curation of HTS data, yielding reliable sets of active and inactive compounds.
• * Demonstration of a practical protocol for data curation and re-uploading.
• * Provision of nine new, high-quality datasets for LB-CADD method development and benchmarking.

Conclusions:

• * A standardized curation process is vital for maximizing the utility of public HTS data for LB-CADD.
• * The provided datasets and protocol will serve as a valuable resource for the scientific community.
• * This work facilitates the development and validation of novel LB-CADD tools for drug discovery.