Toxicity data informatics: supporting a new paradigm for toxicity prediction.

Area of Science:

• Toxicology and Cheminformatics
• Computational Toxicology
• Environmental Health Sciences

Background:

• Chemical toxicity data exists at various descriptive levels, from dose-response to summarized endpoints.
• Predictive toxicology approaches are constrained by the availability and organization of this data.
• Existing data is often fragmented across different sources and formats.

Purpose of the Study:

• To highlight new public data initiatives aimed at unifying chemical toxicity information.
• To explain how data standards facilitate data mining and predictive modeling.
• To demonstrate the role of these initiatives in bridging experimental and computational toxicology.

Main Methods:

• Focus on three key initiatives: ToxML (Toxicology XML standard), DSSTox (Distributed Structure-Searchable Toxicity Database Network), and ACToR (Aggregated Computational Toxicology Resource).
• Emphasize the importance of data standards in enabling data integration and "read-across."
• Describe how "read-across" supports aggregation of data for building predictive models.

Main Results:

• These initiatives are creating a more unified, mineable, and modelable landscape of public toxicity data.
• Data standards are crucial for enabling "read-across" capabilities.
• "Read-across" facilitates flexible data mining and aggregation of toxicity information into modelable endpoints.

Conclusions:

• Shared data standards and aggregation rules are key to integrating diverse toxicity data.
• These efforts span experimental toxicology, computational modeling, and chemical information resources.
• The initiatives effectively bridge the gap between different scientific disciplines and data types for improved predictive toxicology.