Toward a Checklist for Exchange and Interpretation of Data from a Toxicology Study

Jennifer M. Fostel 3 Lyle Burgoon 0 Craig Zwickl 2 Peter Lord 5 J. Christopher Corton 4 Pierre R. Bushel j Michael Cunningham k Liju Fan jk Stephen W. Edwards 4 Susan Hester 4 James Stevens 2 Weida Tong 1 Michael Waters 3 ChiHae Yang 0 Raymond Tennantk 0 Department of Biochemistry and Molecular Biology, Michigan State University , East Lansing, Michigan 48824 1 National Center for Toxicological Research , Jefferson, Arkansas 72079 2 Lilly Research Laboratory , Greenfield, Indiana 46140 3 NIEHS, LMIT ITSS Contract , Research Triangle Park, North Carolina 27709-2233 4 National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency , Research Triangle Park, North Carolina 27711 5 Johnson and Johnson PRD , Raritan, New Jersey 08869 - Data from toxicology and toxicogenomics studies are valuable, and can be combined for meta-analysis using public data repositories such as Chemical Effects in Biological Systems Knowledgebase, ArrayExpress, and Gene Expression Omnibus. In order to fully utilize the data for secondary analysis, it is necessary to have a description of the study and good annotation of the accompanying data. This study annotation permits sophisticated cross-study comparison and analysis, and allows data from comparable subjects to be identified and fully understood. The Minimal Information About a Microarray Experiment Standard was proposed to permit deposition and sharing of microarray data. We propose the first step toward an analogous standard for a toxicogenomics/toxicology study, by describing a checklist of information that best practices would suggest be included with the study data. When the information in this checklist is deposited together with the study data, the checklist information helps the public explore the study data in context of time, or identify data from similarly treated subjects, and also explore/identify potential sources of experimental variability. The proposed checklist summarizes useful information to include when sharing study data for publication, deposition into a database, or electronic exchange with collaborators. It is not a description of how to carry out an experiment, but a definition of how to describe an experiment. It is anticipated that once a toxicology checklist is accepted and put into use, then toxicology databases can be configured to require and output these fields, making it straightforward to annotate data for interpretation by others. This manuscript has been reviewed and approved for publication by the Environmental Protection Agency but does not necessarily reflect the views of the Agency. Mention of trade names or commercial products does not constitute endorsement or recommendations for use. 1 To whom correspondence should be addressed at NIEHS, MD F1-05, PO Box 12233, 111 Alexander Drive, Research Triangle Park, NC, 27709-2233. E-mail: . This article arises from the authors experience with databases, data exchange, and interpretation of cross-study data. It does not describe how to do a study, but rather defines useful information to include in describing the data from the study when it is published or exchanged with collaborators. The checklist herein reflects best practice, i.e., the ideal annotation to include with data to permit interpretation in the context of the study. This checklist focuses on the biological information needed to interpret data from a study, and thus is a logical complement to the technology-focused checklists under development now. With community use, we anticipate that the minimum information required for interpretation of a study will emerge from this checklist. A biological or biomedical investigation is viewed as a selfcontained unit of scientific enquiry. Investigations often include studies of biological subjects, which are examined in situ or in a laboratory, in observational or perturbational studies. This proposal is to expand the data exchange checklist for toxicology/toxicogenomics to reflect the fact that a toxicological study can focus on any one of a number of different subject types. Therefore, it makes more sense to create a checklist tailored to the particular subject type and study design used rather than to aim for a single toxicology study data checklist. This is not intended to be a closed, complete list, but rather to be an illustrative proposal and a living document, so that as additional aspects of study and subject are found to be critical to understanding a study, these pieces will be included in this checklist, and in the data exchange associated with it. Databases such as the CEBS Chemical Effects in Biological Systems Knowledgebase (Waters et al., 2003), ArrayTrack (Tong et al., 2003, 2004), and dbZach (Burgoon et al., 2006) all collect and store study data. At the moment only CEBS is a public data repository, but all three database initiatives will benefit from consensus around the minimal information needed to interpret a study, and therefore from a checklist, such as in this proposal, of the information important to include in a study description. It is important to keep in mind that the data fields included in the following tables are intended to be close to the minimum data required for exchange and interpretation of a biomedical study. The CEBS Data Dictionary (CEBS-DD; Fostel et al., 2005) includes a longer checklist of additional data which enriches interpretation of the study data, and supports meta-analysis of data from multiple studies. The aim of the CEBS-DD was to define the maximal set of data elements that could be used to describe a study; this set is growing as additional studies are deposited in CEBS. The current effort is to identify the minimal set of data elements without which it is difficult or impossible to interpret data from a study. A number of standards and data exchange checklists initiatives are currently underway. At the moment, these initiatives are each focused on a specific technology, but do not fully represent the accompanying biology. Examples include microarray/transcriptomics (the Minimal Information About a Microarray Experiment [MIAME]; Brazma et al., 2001; Microarray Gene Expression Data [MGED] Society Transcriptomics Working Group, http://fugo.sourceforge.net/ community/community.php), proteomics (Protein Standards Initiative [PSI], http://psidev.sourceforge.net/), metabolomics/ metabonomics (Metabolomics Standards Initiative, http://msiworkgroups.sourceforge.net/), in situ hybridization (Minimum Information Specification For In Situ Hybridization and Immunohistochemistry Experiments, http://scgap.systemsbiology. net/standards/misfishie/), etc., or on a particular scientific discipline such as nutrigenomics (MIAME-Nut; see http://www. mged.org/Workgroups/rsbi/rsbi.html) or environmental work (MIAME-Env; http://nebc.nox.ac.uk/miame/miame_env.html). An early effort also created a MIAME-Tox (http://www.ebi. ac.uk/microarray/doc/standards.html). The Minimum Information (...truncated)