Diabetes mellitus (DM) is one of the most common feline diseases seen by veterinarians in the USA. It is caused by the loss or dysfunction of insulin secretion by pancreatic beta cells, diminished insulin sensitivity, or both. The disease requires careful management, including twice daily insulin injections, and the costs of veterinary care and lifestyle changes can significantly impact cat owners. Very little is known about the genetics that contribute to the risk of this complex disease. We have recently identified a significant association with DM, using a genome-wide association study (GWAS) performed on cats genotyped on a custom 340k single nucleotide polymorphism (SNP) Illumina array. We now have the unique opportunity to genotype cats on a custom 2 million SNP Affymetrix array, to perform the densest and largest genome-wide search for loci associated with DM in the domestic cat. We expect to replicate our previous association and identify the variant(s) associated with feline DM using whole-genome sequencing. To complement these results, we will perform RNAseq on pancreas tissue to determine differentially expressed genes between diabetic and healthy cats. This research will enable cats with an increased genetic predisposition for DM to be identified, allowing earlier diagnosis and permitting intervention and monitoring before irreversible disease occurs. Further, cat breeders can use the genetic test to make informed breeding decisions to reduce the frequency of the disease in the population.

We have three specific aims, as follows.

Specific Aim 1: To increase statistical power for GWAS by a) increasing the density of markers queried, from 340,000 SNPs to 2 million SNPs, and b) increasing the sample size, from 67 cases to over 300 cases. Specific Aim 2: To identify variants in the associated regions of interest from aim 1, using whole-genome sequencing (WGS) on ten DM affected and two control cats, in collaboration with the 99 Lives Cat Genome Consortium. Specific Aim 3: To determine pancreatic genes with significantly different expression levels between diabetic and healthy control samples.