toxicology data are generated across many organizations, with the results being reported in the literature or in an online database. In some cases, the full study results may be reported, covering the identity of the tested chemical, the experimental protocol, the study findings (such as details on individual animals treated with histopathology, clinical chemistry, and hematology details recorded), and an overall outcome. In other cases, only summary information on high-level endpoint calls [such as NoAEL or ШАЕЬ (lowest observed adverse effect level ) values or positive/negative/equivocal calls] may be included. the format and level of detail that is collected across the public literature and online databases are highly variable. the experimental protocols, result findings, and description of the tested chemical are also reported using different non-standardized terms. these variables make it difficult to compare different studies.
To support the development and use of new in silico methods, it is highly desirable to organize the chemical and toxicity data in a harmonized manner. A chemical registration and database system alongside a toxicity database where the content is organized in a consistent manner according to well-defined ontologies is necessary to support searching across chemicals for study types and experimental results. It is also essential for in silico model building, building expert alerts systems, and read-across.
The following sections outline several ways to organize toxicity databases in a harmonized manner and summarizes a number of databases covering different toxicity endpoints.