Analysis of commercial and public bioactivity databases

Area of Science:

• Chemoinformatics
• Bioactivity Data Analysis
• Database Merging

Background:

• Bioactivity databases are crucial for chemoinformatics, containing target protein and small molecule binding data.
• Existing databases are often siloed, requiring integration for efficient data retrieval and analysis.

Purpose of the Study:

• To create a unified bioactivity metabase by merging public and commercial databases.
• To assess the data overlap and complementarity between different commercial and public vendors.
• To quantify inconsistencies in activity data across various sources.

Main Methods:

• Standardization of molecular representation, target information, and activity data from vendor databases.
• Development of a single relational database for streamlined data retrieval.
• Comparative analysis of data points from the same scientific articles across multiple vendors.

Main Results:

• Each data source contains unique small molecule activity data due to the citation of different scientific articles.
• Significant inconsistencies were observed when comparing data points derived from the same article across vendors.
• The overlap between commercial and public databases is not complete, indicating complementary data.

Conclusions:

• Merging diverse bioactivity databases into a metabase enhances data accessibility for researchers.
• Despite inconsistencies, utilizing multiple databases remains beneficial due to incomplete data overlap.
• Inconsistencies highlight the need for careful data curation and validation in chemoinformatics.