Environmental Factor, February 2008, National Institute of Environmental Health Sciences
Portier Outlines Strategy for HTS Pathway Analysis
By Eddy Ball
The January 9 talk by NIEHS Associate Director Chris Portier, Ph.D., in Rodbell Auditorium was the first in a series of seminars sponsored by the National Toxicology Program (NTP) Biomolecular Screening Branch. The presentation, "Identifying Toxicity Pathways - Linking Genes, Pathways and Disease," was also the first presentation by NTP of a strategy for systematically analyzing pathways and identifying targets for high-throughput screening (HTS) in toxicology testing, which is an important component of the NTP Roadmap.
Portier, director of the Office of Risk Assessment Research, reported on the results of an exhaustive literature review and multiple database mining effort that took his team over a year to complete. The effort, spearheaded by two postdoctoral fellows working with Portier, Julia Gohlke, Ph.D., and Reuben Thomas, Ph.D., examined the relationships among thousands of genes and several hundred disease phenotypes to produce their "top twenty" list of disease pathways. The data was pulled largely from the 177 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.
According to Portier, the project's scientists are helping to advance HTS toxicology testing in several ways. The team's research succeeded in defining pathways for disease and identifying clusters of related genes, metabolic processes, phenotypes and environmental factors specific to humans. The scientists also identified high-potential signature pathways for the NTP and other HTS investigators, and they explored chemical linkages from the pathways back to diseases.
Moreover, Portier emphasized, the team developed a repertoire of rules and schemes for determining targets in the pathways through both analytical and experimental approaches to pinpoint which of all the activities that are going on at the cellular level should weigh more heavily in pathway linkage analysis (see text box). Their systematic approach promises to help investigators confirm hypotheses as well as generate new ones.
To illustrate the domino-like disease pathway concept, Portier used the example of the BRCA1 gene, a polymorphism of which has been strongly associated with breast cancer risk. "It's not the BRCA1 [polymorphism] itself that's affecting the cancer risk," he explained. "It's how BRCA1 changes the biochemistry of the cell."
"We're looking for things [genes, proteins and other metabolites] where multiple targets in existing pathways are related to the same or a similar disease, and then we choose that pathway as a major target for HTS," he continued. Portier then presented a series of rules for determining which of all the activities that are going on at the cellular level should weigh more heavily in an analysis of the pathways. As Portier explained, investigators need guidelines for selecting the targets that are more clearly related to the disease in question and not ones that are shared with other disease pathways.
Asked about how he plans to validate the predictive ability of HTS findings, Portier proposed comparing the results on chemicals with full animal testing data with HTS pathways analysis of the same chemicals. "If we could show that the chemical, through whatever type of -omics analysis we did, linked to the pathways and then linked to the diseases seen in the animals," he explained, "we'd be in a good position to argue that indeed there is at least some degree of validation going on."
HTS toxicology testing promises to help the NTP and others reduce the use of animals in testing and begin to address the backlog of over 60,000 registered chemicals that have not been thoroughly tested for their possible toxicological, mutagenic or carcinogenic effects But before investigators can realize the potential of HTS with human cell lines, they need to build on the kind of systematic approach Portier outlined for finding the needles in the haystack of millions of target analytes that -omics platforms are capable of identifying.
Selecting the Best Working Targets:
Strategies for Developing a Hierarchy of Importance
Portier offered several different rules and schemes to apply in order to determine the relative importance of targets that show a significant change in relation to disease endpoints:
- Heavy Ends Rule, which gives targets at the edges, the beginning or end of a pathway greater weight than targets in the middle
- Sequential Best Rule, which ranks targets that are adjacent in a pathway more weight than targets with more distant associations
- Crossroads Analysis, which values targets more highly if they appear at the point where two pathways cross each other
- Loner Analysis, which gives more weight to targets that appear only in one pathway associated with a disease
- Watershed Analysis, which manipulates targets by computer or by knock-out to see which targets have the "most flow" downstream in the pathway
- Emerging-disease pathway analysis to link targets in shared pathways with environmental factors
- Constructing phenotype-phenotype, phenotype-environmental factor and environmental factor-environmental factor interaction networks to confirm hypotheses and generate mode-of-action hypotheses linking environmental factors with targets