About the Database
Background
Public health researchers and regulators need scientifically based strategies to screen, and prioritize for testing, the large number of environmental substances that have not been adequately assessed for safety. Some of these substances may pose a risk to the developing fetus. The initial impetus for this Developmental Toxicity database was an ILSI Research Foundation workshop report on the potential value of statistically-based structure-activity relationship (SAR) models as tools for prioritizing chemical substances for further study (see Julien et al., Birth Defects Research Part A 70:902-911, 2004). That report concluded that refined approaches for utilizing in vivo toxicity data would greatly enhance the value of SAR models for developmental toxicity; however, the development of such approaches would require a robust collection of data. In response, ILSI Research Foundation initiated a project to develop a prototype database.
While the initial purpose of the prototype database was to support development of SAR models, the completed product is useful for a range of purposes. The database compiles detailed data on specific developmental and maternal endpoints in a readily accessible electronic format. Its value is enhanced by the guiding principles used to construct the database. Specifically, the ILSI Research Foundation working group aimed to build a database that is:
Objective In place of “summary calls” of authors’ conclusions, this database compiles quantitative dose-response data on discrete endpoints.
Transparent Information on methodology accompanies study results; thus, data may be examined in the context of study design. Full references to the original source of data are provided.
Expert Informed/ Biologically Rational A collaboration between developmental toxicologists and SAR modelers generated the overall database design, which facilitates the aggregation/disaggregation and use of data in a biologically rational way.
Description
The database was constructed using Microsoft ACCESS. It holds “Segment II” type data from published studies, including dose-response data on treatment-related effects for a range of fetal survival endpoints (e.g., changes in corpora lutea, implantations, resorptions, dead fetuses), fetal growth endpoints (weight, anogenital distance), and morphological endpoints (discrete external, skeletal, or visceral abnormalities). Data on maternal effects are also captured, as well as detailed information on the methods used to generate each data set.
Data have been compiled for a range of species (rat, mouse, rabbit, hamster) and routes (oral gavage, feed or water, inhalation, dermal, injection). Version 1.0 holds data from 315 individual experiments, covering 166 publications and 195 different substances. Each substance is identified by its name, Chemical Abstracts Service Registry Number (CASRN) and a Generic Substance ID (GSID). The GSID corresponds to the DSSTox Generic Substance ID used in the EPA’s Distributed Structure-Searchable Toxicity (DSSTox) Public Database Network http://epa.gov/ncct/dsstox/.
For additional information, click on the following links:
Chemicals Included in v1.0 (PDF)
Source Publications for v1.0 (PDF)
Screen Shots of Data Entry Forms (PDF)
Status A full expert review of this database has not been completed; it is being released in its current form to allow broader use, review and comment. Feedback and suggestions are welcome.
Download the Developmental Toxicity Database Version 1.0
Important Note on Macros
This database contains macros to facilitate data sorting, etc. Upon opening the database your machine may alert you to the macros and display an “action failed” message. If so, close the action failed message box, and click on the OPTIONS box next to the security warning. Choose “Enable this content.”
Developmental Toxicity Database v1.0 (.zip file includes Microsoft Access and support files and README PDF)
The “HELP FILE” for the database is undergoing revision. To be notified about the new HELP file and other updates and corrections, please send your email address to:DevToxDB@ilsi.org
For more information: please contact Beth Julien (bjulien@LifeSciResearch.com) or Rick Canady (rcanady@ilsi.org).
Overview of Methods
Database Structure: Data Fields and Vocabulary
Decisions regarding the specific types of data to collect, the vocabulary to use, and the level of detail to capture were made during cross-disciplinary meetings of toxicologists, chemists, database experts and potential database users. These meetings generated the core database structure, which was modified during the data harvest process only as necessary to allow consistent and transparent data collection. Internationally harmonized terminology for teratogenic effects (“Terminology of Developmental Abnormalities in Common Laboratory Mammals” Wise et al. Teratology 55:249-292, 1997) was incorporated into the database design.
Criteria for Study Selection
Version 1.0 contains data from studies published in peer-reviewed journals. Studies were identified using a search strategy developed in consultation with staff of the National Library of Medicine (NLM), and focused first on the NLM’s Developmental and Reproductive Toxicology Database. Studies were generally excluded if: i) the test substance was a mixture, plant extract, emulsion, slurry; ii) the treatment period was pre-mating, or included treatment of males; iii) the test species was not standard; iv) the publication was not in English; or v) the publication contained very little or no original data. Due to limited resources, not all suitable studies could be entered into the database; however, there was no systematic attempt to prioritize studies based on study completeness or quality. The working group recognized that studies should be filtered by database users at the point of data retrieval. Filtering studies at this point (as opposed to filtering upon data entry) maintains database transparency and flexibility for a range of uses. The ILSI Research Foundation working group recommended that, as the database grows, expert guidance be developed to facilitate its use.
Data Harvest and QC
In line with the overall project aim of an objective and transparent database, the team responsible for data review and harvest developed a set of “rules” to deal with the wide range of study designs encountered in published studies. This rules list, along with frequent team teleconferences, promoted consistency in the data harvest process (consistency across publications and across personnel entering data).
Complete List of Rules
Acknowledgements
ILSI Research Foundation thanks the project’s financial sponsors, as well as the many individuals who gave their time and expertise to produce this database.
Financial Sponsors
Funding for design and construction of the database, and for initial data review and harvest was provided by the Health Canada Existing Substances Division and the EPA Office of Pollution Prevention and Toxics. Funding for review of entered data was provided by the Health and Environmental Sciences Institute (HESI) Developmental and Reproductive Toxicology (DART) committee.
Scientific Design
Chihae Yang, with help from Mary Shackelford and Yan Gu, proposed the initial database structure, which was fundamental to the effort and initiated a highly productive cross-disciplinary discussion. A core group of developmental experts worked countless hours to debate, test, and refine the database concepts and design: Dave Wise, John DeSesso, Paul Foster, Jennifer Seed, Calvin Willhite, Don Stump, and Tom Collins. Phil Mirkes and Nigel Brown were also instrumental in developing the original database concepts, and providing advice on early design. Expertise in database development and computational approaches was provided by Chihae Yang, Ann Richard, Ed Matthews and Dan Benz. Ann Richard played an essential role, clearly explaining the perspective and needs of modelers and chemists. Bette Meek provided critical expertise from the perspective of the regulatory community. Beth Julien led the overall process of integrating and reconciling the perspectives of the various scientific disciplines. She also coordinated the database construction, data harvest and review process. Many experts provided critical feedback on early drafts of the database, testing its suitability with case examples from the literature, notably: Bill Breslin, Neil Chernoff, Bob Clark, Donna Farmer, Yan Gu, Barbara Hales, Deborah Hansen, Kok-Wah Hew, Bob Kavlock, Mary Schakelford and Mary Kate Ziejewski. A method for organizing entry of clinical signs data was adapted from a system provided for use in this project by Alan Hoberman, Charles River Laboratories.
ILSI Research Foundation also acknowledges the following consultants who implemented the scientific design.
Cynthia Van Landingham, ENVIRON International Corp., translated the expert working group’s design concepts into an ACCESS database. Data entry rule development, data harvest, and data QC tasks were accomplished by a team of scientists with extensive professional experience in the field: Judy Stevens, Annette Iannucci, Maureen Urbaniak, Gloria Jahnke and Judy Buelke-Sam. Special thanks are due to Judy Stevens and Annette Iannucci for their meticulous work on the final data review and rule development process.