Studies in model organisms can help elucidate the molecular mechanisms causing disease, and the dog is an excellent example as it shares many of the same environment, diseases and genes with humans. With mapping tools available, we are in a unique position to map canine disease genes to learn about human biology and medicine. Using the same principle as for our previous study of 39 tumor/normal pairs (T/N) from human glioblastoma patients, we will look for non-coding constraint mutations (NCCMs) in multiple canine cancers. These studies will implicate cancer driver genes based on changes in regulatory elements. Identification of NCCMs in canine osteosarcoma (OSA) and lymphoma (LSA) We have previously generated OSA T/N whole-exome sequencing (WES) data from three breeds, thereby identifying significantly mutated genes similar to human OSA. Also, we found frequent mutations in SETD2, a methyltransferase gene and known tumor suppressor. We have now generated T/N whole genome sequencing (WGS) data from 93 dogs with osteosarcoma and combined this with data from tumor normal pairs from a public repository. We will use this data to detect both protein-coding mutations and NCCMs. Categorizing individuals by presence or absence of protein-coding mutations in SETD2, as well as potential regulatory mutations as NCCMs near SETD2, we will select three groups of samples and perform methylation analysis with the aim of identifying differentially methylated regions resulting from dysfunction of SETD2. Canine multicentric b-cell lymphoma For canine multicentric b-cell lymphoma we and collaborators have now performed WGS of 108 T/N pairs. In a previous WES study, we found that the mutational spectrum of mutations overlapped between breeds. Based on this novel data we will see if NCCMs will allow us to subcategorize the tumors on a broader scale. A total of 50 T/N pairs have been included in a clinical trial with multiple arms and for a subset of samples we have T/N WGS from both treatment naïve samples and samples taken at the time of disease relapse. While our power may not be enough, we will still try to see if NCCMs in specific genes may somehow be related to treatment outcome and tumor resistance. For both these tumor types we will compare known and novel driver genes with those found in human samples of the same phenotypes, OSA and LSA, in the International Cancer Genomics Consortium (ICGC) data set.