Primary central nervous system tumors comprise more than 100 histologically distinct tumor types, some with extremely poor prognoses. A major challenge in the advancement of brain tumor treatment is the difficulty of obtaining a definitive diagnosis prior to initiating treatment. As up to 10% of stereotactic biopsy procedures are associated with intercranial bleeding, biopsies are often delayed until debulking surgery. Evidently, there is a critical need for improved diagnostic modalities and therapeutic regimens. The dog has been recognized as a promising animal model for brain tumors because pet dogs develop spontaneous brain tumors with intact immuno-microenvironments with a relatively high incidence. This research project has two aims 1) To identify ways of diagnosing canine brain tumors without the need for invasive biopsies. This is done by evaluating the mutational landscape in cfDNA from cerebrospinal fluid from dogs with brain tumors and comparing this to mutations identified within tumor tissue from the same individuals, with the view that this could allow dogs to enter comparative clinical trials in the future with a more secure diagnosis. 2) To evaluate non coding constraint mutations in canine glioma samples and compare these to data from humans which is in the process of being published and hence will be available for comparison. We have generated new WGS data from 15 brain tumors from which we also have cfDNA to use for aim 1 of the project. For aim 2 we will use data from 67 WGS glioma/normal pairs from dogs which is available to download and realign this on our novel canine reference genome CanFam 4.0.