Project 2

Toxico-Genetic Modeling: Population-Wide Predictions from Toxicity Profiling

Description

Project 2 is focused exclusively on the development of tools for the analysis of the toxicological data in genetically diverse subjects by taking into account the information on high-density (up to millions of SNPs) genetic maps of the test populations. The results will be used to provide toxico-genetic context to the development of mechanistic, predictive, and statistical models in Projects 1 and 3.

The inference and modeling proposed in Projects 1 and 3 follow well-established approaches for integrating multiple data sources with a goal to:

a) elucidate the high-level topology of regulatory networks induced in response to chemical perturbation;

b) estimate rate parameters and other quantities in mechanistic temporal pathway models;

c) use and deploy software for computational prediction of response to chemical perturbation; and

d) develop validated and predictive Quantitative Structure-Toxicity Relationship models that employ both chemical and biological descriptors of molecular structures and take into account genetic diversity between individuals.

Although numerous existing data sources will be used (and are described in Projects 1 and 3), gaps remain in key areas of toxicological relevance. In particular, variation in inherited genetic susceptibility is known to play a large role in toxicological response. Datasets based on lower organisms (such as yeast), or using only a single cell line/rodent strain, cannot be used to understand how constitutional genetic variation between humans, or other organisms, affects toxicity. Datasets which combine global expression profiling and high-density genotyping, along with traditional toxicity data, enable the discovery of relevant susceptibility variation. In addition, genetic variation in toxicity response provides evidence for causation and proximal influence that is difficult to obtain using other omics technologies. Genetic variation causes network perturbations that can help to resolve the activity of transcription factors in an unbiased manner, without relying on the prior accuracy of assumed transcriptional regulation.

As -omics and SNP genotyping technologies become less expensive and more standardized, the foremost challenge lies in analyzing such data, data that will be inevitably a part of standard toxicological research in the very near future. To effectively handle the massive datasets, we have assembled an experienced team and a set of novel analytic approaches. Project 2 benefits from a convergence of areas of research excellence at UNC to address various facets of expression quantitative trait locus (eQTL) analysis. The approaches can be roughly divided as follows:

a) extremely fast computation of genotype-phenotype correlations;

b) deep statistical investigations of the transcriptional program underlying expression variation; and

c) hypothesis testing combining expression and sequence data to identify pathways related to toxic response.

A targeted series of new biological experiments (representing a very small portion of the overall budget) will be conducted to gather new toxico-genetic data. These novel data will be combined with other public data (such as SNP maps of mouse and man) to elucidate and dissect the genetic variation of toxicity response in mouse and human cells. The results will be used to further inform the outputs of Projects 1 and 3. As an ancillary effort, Project 2 will systematically analyze these data and public datasets to determine the relevance genetic polymorphism data to known pathways in the toxicity response. The new software created as part of the Project’s Specific Objectives will become a valuable part of the toolbox of genetic modelers and other toxicological researchers.

This project pursues the following major specific objectives:

• Develop toxicogenetic expression Quantitative Trait Loci (eQTL) mapping tools, perform transcription factor network inference and integrative pathway assessment
• Perform toxicogenetic modeling of liver toxicity in cultured mouse hepatocytes
• Discover chemical-induced regulatory networks using population-based toxicity phenotyping in human cells