A comprehensive bioinformatics approach toward the molecular characterization of vertebrate limb specification and development.

The developing vertebrate limb is an ideal model system for investigating the genetic mechanisms underlying the morphological and evolutionary diversity between serially homologous structures. In the developing embryo, limb-type specific expression of Pitx1, Tbx4, and Tbx5 has been shown to partially regulate the determination of limb identity. However, numerous lines of evidence, including the fact that these three genes encode transcription factors, suggest that additional candidate limb identity selector genes exist. In order to decipher the complex network of genetic interactions regulating limb morphology, Serial Analysis of Gene Expression (SAGE) was used to generate a comprehensive, unbiased and quantitative gene expression profile from populations of mRNA purified from mouse E11.5 forelimbs and hindlimbs. 68,406 forelimb and 68,450 hindlimb SAGE Tags representing 36,300 unique transcripts were sequenced. A program called eSAGE was written to manage and analyze the SAGE data. Additional computational tools and eSAGE facilitated the identification and regulation of a critical variable in the SAGE protocol that has the potential to bias a population of SAGE Tags relative to the GC content of the SAGE Tag DNA sequence. Using eSAGE to generate and compare these high-quality gene expression profiles, a set of candidate genes predominantly expressed in the forelimb or hindlimb was identified that included Pitx1, Tbx4, and Tbx5. This set of genes was refined by performing a 'virtual' subtraction with 362,344 Tags from 6 additional non-limb SAGE libraries, reducing the number of candidate genes by 74%. The ehm-Tag-Mapping method was developed to accurately identify UniGene clusters represented by a given SAGE Tag. A subset of these candidate genes are currently being evaluated by semi-quantitative RT-PCR and whole mount in-situ hybridization. Since 84% of the 36,300 unique transcripts were expressed on average 1 copy per cell or less, statistical simulations were performed to determine the quantitative accuracy of low abundance. This thesis provides a molecular framework for exploring the evolutionary divergence of morphology and function between serially homologous structures. Furthermore, the bioinformatics tools and the overall strategy used here are broadly applicable to other SAGE-based studies in a variety of organisms.Ph.D.Biological SciencesGeneticsUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/126022/2/3016914.pd