You are here:
Utilizing automated and semi-automated data analytic tools for curating data in the ECOTOX Knowledgebase
Citation:
Hoff, D., J. Olker, C. Elonen, A. Pomplun, A. Anderson, A. Pilli, C. Suomi, K. Nehiba, AND T. Gephart. Utilizing automated and semi-automated data analytic tools for curating data in the ECOTOX Knowledgebase. SETAC North America, Fort Worth, TX, November 15 - 19, 2020. https://doi.org/10.23645/epacomptox.13241546
Impact/Purpose:
The ECOTOX Knowledgebase (www.epa.gov/ecotox) is the world’s largest compilation of ecotoxicity data and has provided support for chemical assessments and research for over 30 years through systematic and transparent procedures to identify and curate ecological toxicity data. Since its inception, the number of chemicals introduced into commerce has, and continues, to grow. With an ever-increasing number of new chemicals coupled with goals of reduction in animal usage for test
Description:
The ECOTOX Knowledgebase (www.epa.gov/ecotox) is the world’s largest compilation of ecotoxicity data and has provided support for chemical assessments and research for over 30 years through systematic and transparent procedures to identify and curate ecological toxicity data. Since its inception, the number of chemicals introduced into commerce has, and continues, to grow. With an ever-increasing number of new chemicals coupled with goals of reduction in animal usage for testing, there is a tremendous need for streamlined processes to curate high volumes of good quality existing information within shorter timeframes. Through the decades, ECOTOX operating procedures have been developed which are consistent with attributes of standardized systematic review protocols. Most steps of those processes were highly dependent on manual curation, often considered the gold standard for accuracy and precision, but very labor and time intensive. To address the challenges of curating data for more chemicals, more rapidly, well established ECOTOX protocols for manual curation have evolved to include automated and semi-automated data analytic tools and techniques to streamline processes. These include: 1) generation of chemical search terms; 2) automated searches from literature inventories to retrieve citation data; 3) title and abstract screening; and 4) generation of data evaluation records. To date, these new processes have demonstrated a reduction of up to 83% in level of effort to identify relevant journal articles through title and abstract screening steps in the systematic review process. These increases in efficiencies have not sacrificed accuracy in identifying pertinent literature. Comparisons to date between literature identified through manual curation steps versus through new data analytic tools have been found to be very consistent. In the past, ECOTOX has focused curation efforts by prioritizing targeted chemical searches requested by EPA Offices conducting assessments but has also endeavored to identify current ecotoxicity content from searches of journals that contain pertinent ecotoxicity data. The data analytic tools currently being employed, as well as others under development, will allow ECOTOX to continue to be responsive to chemical requests from regulatory programs while concurrently increasing the amount of current content extracted for chemicals of broader interest in the ecotoxicological community.
URLs/Downloads:
DOI: Utilizing automated and semi-automated data analytic tools for curating data in the ECOTOX Knowledgebase