Here I report my collaborative work employing genome editing to generate mutations in diverse model organisms to accelerate reverse genetic studies. We confirmed the mapping of a modifier of endogenous retroviral mutations with nucleotide-level precision by creating a point mutation in Nxf1 which mimics a naturally occurring, suppressive allele. I also applied this technology to a novel, highly conserved protein coding gene, nmf9/Ankfn1. I confirmed the positional cloning of this gene, identified in a forward genetic screen for vestibular behaviors in mice, by creating alleles which phenocopied the behavioral outcomes of the ENU-induced mutation. We validated functional conservation with mutant alleles of the Drosophila homolog, disrupting two annotated functional domains and an additional conserved, unannotated domain. Mutant flies had locomotor defects, abnormal sleep patterns, and early lethality. The quick adoption, refinement, and distribution of methods for making transgenic animals facilitated other collaborative projects. In one, I created an amino acid substitution in ELOVL2 to selectively inactivate its ability to process certain polyunsaturated fats important for retinal and brain function. This mutation recapitulated molecular signatures of aging in the murine retina predicted by progressive loss of function associated with methylation of the gene promoter. In another, I provided methodological insight to dissect an enhancer element involved in monocyte development. Enhancer deletions in the Nr4a1 gene identified sequences which alter Ly6Clow monocyte development. Chapter 3 reports my study of Ankfn1 zebrafish, where I saw the opportunity to test functional conservation of Ankfn1 paralogs. Zebrafish have two copies of this gene, an ancestral copy which contains a Ras-association (RA) domain, and a derived copy which does not. I analyzed the physicochemical constraint of 115 homologs of Ankfn1 representing 800 million years of evolutionary divergence and edited the most constrained peptide sequence to create disrupting alleles. Interestingly, I discovered an overt swim bladder phenotype which was, with two exceptions, observed only in fish with mutations in both alleles of each paralogous copy. These studies expand our understanding of a genetic modifier of IAP insertions, characterize a gene important in neurological function, and demonstrate the power of gene editing to accelerate reverse genetics.

In the business world today there is a vast amount of information, and in order to process this information it must be structured. I develop a set of classifiers and corresponding user interfaces to assign tags to data that correspond with a structured framework. The classifiers are applied through a case study in the support area of a major networking company. The three classifiers provide tags for technical support documents with an F measure of .585, customer service requests with an F measure of .706, and customer support forum messages with an F measure of .78.