DISCOVERY OF CANINE CANCER-ASSOCIATED SOMATIC MUTATIONS IN WES, WGS, AND RNA-SEQ

Spontaneous cancers in dogs provide valuable yet underutilized models for cancer research. To enhance their utility, we comprehensively analyzed DNA sequencing data from 684 canine tumors using whole exome sequencing across over 35 breeds and 7 common cancer types. Our results demonstrated that the genetic landscape of canine cancers is predominantly driven by tumor type, with each cancer displaying distinct mutational patterns that mirror those in human counterparts. For example, canine mammary tumors had frequent PI3K pathway mutations, while osteosarcomas showed a high prevalence of TP53 mutations. We also found variable tumor mutation rates across cancer types, with higher rates in oral melanoma, osteosarcoma, and hemangiosarcoma but lower rates in mammary tumors and gliomas. Interestingly, mutation rates are consistently associated with TP53 but not PIK3CA mutations. In parallel, we developed an efficient pipeline to identify somatic mutations from tumor-only RNA-seq data by leveraging a large database of known human cancer mutations, variant allele frequencies, and machine learning. This approach reduced the need for matched normal samples while recapitulating expected mutation patterns, such as PIK3CA mutations in mammary tumors, and revealing consistent mutation rate patterns across canine cancer types. Our integrated analyses provide optimized methods to unlock the genetics of spontaneous canine cancers for translational insights into human disease.