Bioinformatic approaches to facilitate canine disease genetics and genomics

Abstract

Since the annotation of the dog genome in 2005, dogs have emerged as excellent models of human disease. Many disease associations of variant alleles in homologous genes have been discovered in dogs, providing new therapeutic candidates to the corresponding human diseases as well as establishing preclinical large animal models. Significant progress in genetic studies has happened after moving from microarrays to next generation sequencing or combining the two approaches. This transition has required the development and application of novel bioinformatic approaches to facilitate genetics and genomics. This thesis established a variety of bioinformatic approaches and tools to facilitate canine genomics and disease gene discovery.

In study I, the successful development and application of the bioinformatic pipelines resulted in the identification of causal variants of three new disease genes, SLC37A2, SCARF2 and FAM20C, of relevance to Caffey disease, van den Ende Gupta syndrome (VDEGS) and Raine syndrome in human, respectively.

In study II, novel genomic content was discovered through de novo assembly of genomic reads of Border Collies, which didn’t map to the current canine genome reference. This study revealed sequences that filled the existing gaps in the reference genome and identified gene models that were missing from the reference. Overall, this study reveals novel genomic content to facilitate the improvement of upcoming canine genome reference for disease variant allele discovery in candidate genes.

In study III, a novel bioinformatic tool, webGQT, was successfully developed and piloted to handle and filter large amounts of next generation sequencing (NGS) data. This tool is purported to non-bioinformatics users to mine genetic information from millions to billions of variants among thousands of genomes. This tool has been successfully utilized in various disease genetics projects.

In summary, new bioinformatic approaches have been successfully developed and applied in this thesis to facilitate both the transition of the field to the NGS era and disease gene discovery and genomics in dogs. These findings in this thesis have implications to veterinary research, diagnostics and human medicine with novel candidate genes in three rare disorders.