Genotype-Phenotype Relationships and the Patterning of Complex Traits as Exemplified in the Mammalian Dentition

The relationship between genes and traits is complex, involving the interaction among genes, environment, and chance. However, common practice in genetics treats this relationship as a straightforward one-to-one mapping from genotype to phenotype. The roots of this practice can be traced to Mendel who chose traits with a direct relationship between genetic variation and phenotypic variation in formulating his particulate theory of inheritance. It has been further solidified by the successes of modern genetics in identifying genes involved in many simple traits, such as rare human diseases. However, most traits are not simple and to understand complex traits it is necessary to decipher the developmental processes that occur between genes and traits. To date most attempts to build developmental models of traits have relied on natural or experimentally induced mutations that produce pathologic-scale phenotypic effects, but the raw material of most evolution is normal variation in traits found among members of a population. The goals of this dissertation, therefore, are to describe the cellular mechanisms and developmental processes that lead to a many-to-many genotype-phenotype relationship and to attempt to decipher the genetic and developmental processes that produce normal variation in traits using the mammalian dentition as a model system. Two approaches to understanding the genotype-phenotype relationship are described and examples given of how both lead to a many-to-many relationship. First, cellular and genetic mechanisms, such as alternative splicing, DNA and chromatin modification, cellular gene choice, and gene regulation, which lead from DNA sequence to protein structure, are discussed. And, second, examples of variation in the genotype-phenotype relationship which can produce variable phenotypes from the same genetic information and stable phenotypes despite genetic variation are presented. To examine how normal variation in complex repeated traits such as the mammalian dentition is produced two experimental approaches are taken. First, the AXB/BXA recombinant inbred mouse set is used to genetically map normal variation identified in cusp pattern between the teeth of C57BL/6J and A/J mice. Three chromosomal regions on mouse chromosomes 11, 13, and 19 reach the suggestive level of association with the trait. Within these regions candidate genes are identified by direct sequence comparison between the C57BL/6J and A/J mice. Two genes, Itga3 and Glis3, with non-synonymous substitutions and several genes with potential regulatory variation, including epiprofin/Sp6, may play a role in producing the variation identified. Additionally, identification of all genes in the mapping regions that are expressed in developing teeth using the GenePaint database suggests that traditional methods of identifying candidate genes are biased. In the second approach the expression of the signaling molecule Bmp4 was altered transgenically using the epithelium-specific enhancer of the Dlx2 gene. No overt phenotype was produced suggesting that tooth development is robust to over-expression of Bmp4. Examination of gene expression of downstream targets and antagonists of BMP4 suggests that a complex buffering mechanism is involved. This dissertation highlights the complexities of the genotype-phenotype relationship and attempts to directly confront this complexity by developing methods to study normal variation in traits. The question of how normal variation is produced is complex and no simple answers are provided, but the importance of explaining complex common traits and the production of normal variation for biomedical purposes and our basic...