TOXICOLOGICAL HIGHLIGHT |
Putting the Fun Into Functional Toxicogenomics
Laboratory of Pharmacology and Chemistry, National Institute of Environmental Sciences, Research Triangle Park, North Carolina 27709
1 For correspondence via e-mail: cunning1{at}niehs.nih.gov.
Received May 26, 2006; accepted May 30, 2006
The term "toxicogenomics" was first coined in 1999 to describe the marriage of toxicology and genomics (Nuwaysir et al., 1999
). Since then, the field of toxicogenomics has undergone a rapid and uneven surge of growth. Driven by the promise of whole-genome gene expression analysis by microarray, toxicogenomics has been advanced as the tool for improved mechanistic toxicology screens, more sensitive and earlier toxicity discovery, drug and chemical safety assessments, and new drug discovery assays (Hayes and Bradfield, 2005). However, it quickly became evident to scientists conducting toxicogenomic studies that making sense of enormous data sets from microarray studies would require new tools and new approaches in order to organize and interpret them. Furthermore, the ultimate goal of understanding the integration of known toxicological processes with the information provided by the new "-omic" technologies seemed unfathomable.
The early years of the microarray era were technology driven resulting in homemade and custom arrays with limited usefulness. They had relatively few genes on them and little quality control, and cross-laboratory research was not possible. During this time, the comments during presentations went along the line of "I can do a microarray study but don't ask me what it means." This led to the need to integrate genomic data with a traditional toxicological effect, such as histopathological, clinical chemistry, or ultrastructure analyses. This "phenotypic anchoring" concept was a step forward to understanding toxicological cause and genomic effect and resulted in a more basic understanding of metabolic pathways interrupted and pathological sequelae following toxicant exposure (Paules, 2003). Comments at presentations sounded like "I know how many genes were altered by the treatment and they have something to do with some biochemical pathways I am vaguely familiar with."
The latest approach to data analysis of these huge whole-genome data sets ushered in the concept of functional genomics, which has many interpretations for what the term really means. A long-standing definition of functional genomics is "the development and application of global (genome-wide or system-wide) experimental approaches to assess gene function by making use of the information and reagents provided by structural genomics [in the original more limited sense of construction of high-resolution genetic, physical and transcript maps of an organism]. It is characterized by high throughput or large-scale experimental methodologies combined with statistical and computational analysis of the results. The fundamental strategy is to expand the scope of biological investigation from studying single genes or proteins to studying all genes or proteins at once in a systematic fashion (Waters and Fostel, 2004). Functional genomics aims to discover the biological function of particular genes and to uncover how sets of genes and their products work together in health and disease. In its broadest definition, functional genomics encompasses many traditional molecular genetic and other biological approaches" (Hieter and Boguski, 1997). Which, if you are still with me here, leads to the article in this issue selected for the Toxicological Highlight (Yu et al., 2006).
Yu et al. (2006) have developed a systematic approach to quantitate the degree to which functional gene systems change across dose or time course, linking functional gene category results with the original gene expression data. This program termed GO-Quant is proposed to provide the user a wider quantitative analysis that may be applied to risk assessment.
To put this approach into perspective, consider how data analysis evolved for microarray data sets that may contain 12–20,000 or more genes. Initially, researchers identified statistically significant gene changes and cross-referenced databases, such as National Center for Biotechnology Information, Panther, or MAPPFinder, to understand which biological pathways each gene participated in. Sets of genes could then be assigned to define which function was altered by the changes induced by the treatment (Heinloth et al., 2004). This was a time-consuming approach that was aided in part by the efforts of the Gene Ontology (GO) Consortium (www.geneontology.org). The GO Consortium standardized annotation terms that advanced the use of databases to identify biochemical pathways and processes that resulted in the gene changes detected on a microarray. Prior to the GO Consortium, searching across different databases was often restricted by the search terms used to index the data. One database may use the term protein synthesis and another the term translation, and there would be confusion between the results of a search even though these are functionally equivalent terms. To address this problem, the GO Consortium members have developed three vocabularies (ontologies) to describe gene products as a function of their biological processes, their cellular components, and their molecular/biochemical functions. Simply put, GO describes how gene products function within the cell. The biological process function describes the broad biological context within which the gene product of interest participates, for instance, DNA repair. There will be multiple pathways within a biological process, and it comprises a big picture function. The cellular component is the organelle within which the gene product functions (nucleus). The molecular function describes the catalytic activity of the gene product (base-excision repair) (see Currie et al., 2005).
MAPPFinder works with GO to identify global molecular/biochemical trends in gene expression data (http://www.genmapp.org/MAPPFinder.html). From a set of microarray data, MAPPFinder calculates the percentage of gene changes by the treatment according to the biological process, cellular component, and molecular function described by the GO term above. This allows the user to identify how completely a biological process is affected compared to other biological processes based on gene expression data, hopefully providing the user some understanding of the complete biological effects produced by the treatment. GO combined with pathway mapping can therefore be useful tools to analyze complex gene expression changes induced by a toxicant across different microarray platforms and between laboratories (Bammler et al., 2005).
However, this approach was not intended to analyze whether dose- or time-dependent changes in gene expression needed risk assessment, and the current paper highlighted in this issue addresses this issue with the development of a new program (GO-Quant) to facilitate toxicogenomic data into risk assessments. This program is a system-based one that automatically links functional gene category analysis and calculates the average intensity or ratio for each gene exhibiting statistically significant alteration following toxicant exposure. This approach provides a quantitative analysis to examine gene expression changes across dose and time. Proof-of-principle utilization of GO-Quant is demonstrated by analyzing a database containing gene expression data in rats following exposure to sulfur mustard in a dose- and time-dependent fashion. Their results demonstrated the utility of using GO-Quant to analyze complicated data sets to provide quantitative measurements of global gene expression in a way that can be tailored to the needs of risk assessments.
REFERENCES
Bammler, T., Beyer, R. P., Bhattacharya, S., Boorman, G., Boyles, A., Bradford, B., Bumgarner, R., Bushel, P. R., Chaturvedi, K., Choi, D., et al. (2005). Standardizing global gene expression analysis between laboratories and across platforms. Nat. Methods 2, 351–356.[CrossRef][Web of Science][Medline]
Currie, R. A., Bombail, V., Oliver, J. D., Moore, D. J., Lim, F. L., Gwilliam, V., Kimber, I., Chipman, K., Moogs, J. G., and Orphanides, G. (2005). Gene ontology mapping as an unbiased method for identifying molecular pathways and processes affected by toxicant exposure: Application to acute effects caused by the rodent non-genotoxic carcinogen diethylhexylphthalate. Toxicol. Sci. 86, 453–469.
Hayes, K. R., and Bradfield, C. A. (2005). Advances in toxicogenomics. Chem. Res. Toxicol. 18, 403–414.[CrossRef][Web of Science][Medline]
Heinloth, A. N., Irwin, R. D., Boorman, G. A., Nettesheim, P., Fannin, R. D., Sieber, S. O., Snell, M. L., Tucker, C. J., Li, L., Travlos, G. S., et al. (2004). Gene expression profiling of rat liver predicts potential adverse effects. Toxicol. Sci. 80, 193–202.
Hieter, P., and Boguski, M. (1997). Functional genomics: It's all how you read it. Science 278, 601–602.
Nuwaysir, E. F., Bittner, M., Trent, J., Barrett, J. C., and Afshari, C. A. (1999). Microarrays and toxicology: The advent of toxicogenomics. Mol. Carcinog. 24, 153–159.[CrossRef][Web of Science][Medline]
Paules, R. (2003). Phenotypic anchoring: Linking cause and effect. Environ. Health Perspect. 111, A338–A339.[Web of Science][Medline]
Waters, M. D., and Fostel, J. M. (2004). Toxicogenomics and systems toxicology: Aims and prospects. Nat. Rev. 5, 936–948.[CrossRef]
Yu, X., Griffith, B., Hanspers, K., Dillman, J. F., III, Ong, H., Vredevoogd, M. A., and Faustman, E. M. (2006). A system based approach to interpret dose and time-dependant microarray data: Quantitative integration of GO ontology analysis for risk assessment. Toxicol. Sci. (in press).
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||