INTRODUCTION: Dogs develop many cancers spontaneously and share environmental exposures, immune systems, and disease features with humans. However, canine tumor sequencing studies were generated by different laboratories with different methods, making direct comparisons difficult. A standardized genomic resource can therefore improve reuse of existing data and strengthen comparative cancer research. The Canine Cancer Genome Atlas (CCGA) addresses this need by harmonizing sequencing data across 10 tumor histotypes within a reproducible analytical framework. RATIONALE: We built the CCGA by reanalyzing public sequencing data from the protein-coding regions of 667 tumors and matched normal samples across 10 canine cancer types and >40 breeds. All datasets were processed with common quality control and workflows for mutation calling and detection of chromosome-copy changes, large structural changes, and mutation patterns. We also harmonized available clinical metadata; assessed tumor mutational burden, candidate driver genes, and pathway alterations; and developed an interactive portal to support data exploration, cross-study comparisons, and reuse. Conclusions were calibrated to cohort size, and five well-sampled canine cancers were compared directly with corresponding human cancers. RESULTS: The harmonized analysis identified 44,192 acquired single-base changes, 2272 small insertions or deletions, and 39 genes whose mutation patterns were consistent with positive selection in canine cancers. TP53 was the most broadly altered gene. Recurrent chromosome-copy changes included amplification of MYC and deletion of PTEN, and mutation burden varied markedly among cancer types. Osteosarcoma showed the highest median proportion of the genome affected by copy number aberrations (45.7%). DNA repair pathway alterations were most frequent in hemangiosarcoma (73%) and osteosarcoma (67%). Evidence of positive selection also implicated MAP3K14 and SETD2, with primary inference based on well-sampled cohorts. Direct dog-human comparisons showed shared and species-enriched changes in genes, biological pathways, mutation patterns, and chromosome-copy architecture in melanoma, osteosarcoma, B cell lymphoma, mammary carcinoma, and hemangiosarcoma. The accompanying public portal provides downloadable datasets and interactive tools for exploration across studies (https://caninecancergenomeatlas.org/). CONCLUSION: By reprocessing heterogeneous studies through one analytical framework, the CCGA converts dispersed sequencing data into a reusable comparative resource. It distinguishes robust findings from observations limited by small cohorts and provides a practical foundation for studying cancer mechanisms in dogs, prioritizing hypotheses for veterinary precision oncology, and connecting naturally occurring canine cancers with human disease. Shared genomic alterations and direct comparisons between species support investigation of conserved oncogenic mechanisms, while downloadable data and interactive tools facilitate exploration and reuse. However, incomplete clinical annotation and uneven cohort representation limit clinical interpretation; and therapeutic hypotheses require further validation.
Mazzone, E., Marconato, L., Fanelli, A., Cascione, L., Bertoni, F., Marchiò, C., et al. (2026). The genetic landscape of canine tumors: Insights from the Canine Cancer Genome Atlas (CCGA). SCIENCE, 394(6825), 739-754 [10.1126/science.aeg8777].
The genetic landscape of canine tumors: Insights from the Canine Cancer Genome Atlas (CCGA)
Marconato, Laura;
2026
Abstract
INTRODUCTION: Dogs develop many cancers spontaneously and share environmental exposures, immune systems, and disease features with humans. However, canine tumor sequencing studies were generated by different laboratories with different methods, making direct comparisons difficult. A standardized genomic resource can therefore improve reuse of existing data and strengthen comparative cancer research. The Canine Cancer Genome Atlas (CCGA) addresses this need by harmonizing sequencing data across 10 tumor histotypes within a reproducible analytical framework. RATIONALE: We built the CCGA by reanalyzing public sequencing data from the protein-coding regions of 667 tumors and matched normal samples across 10 canine cancer types and >40 breeds. All datasets were processed with common quality control and workflows for mutation calling and detection of chromosome-copy changes, large structural changes, and mutation patterns. We also harmonized available clinical metadata; assessed tumor mutational burden, candidate driver genes, and pathway alterations; and developed an interactive portal to support data exploration, cross-study comparisons, and reuse. Conclusions were calibrated to cohort size, and five well-sampled canine cancers were compared directly with corresponding human cancers. RESULTS: The harmonized analysis identified 44,192 acquired single-base changes, 2272 small insertions or deletions, and 39 genes whose mutation patterns were consistent with positive selection in canine cancers. TP53 was the most broadly altered gene. Recurrent chromosome-copy changes included amplification of MYC and deletion of PTEN, and mutation burden varied markedly among cancer types. Osteosarcoma showed the highest median proportion of the genome affected by copy number aberrations (45.7%). DNA repair pathway alterations were most frequent in hemangiosarcoma (73%) and osteosarcoma (67%). Evidence of positive selection also implicated MAP3K14 and SETD2, with primary inference based on well-sampled cohorts. Direct dog-human comparisons showed shared and species-enriched changes in genes, biological pathways, mutation patterns, and chromosome-copy architecture in melanoma, osteosarcoma, B cell lymphoma, mammary carcinoma, and hemangiosarcoma. The accompanying public portal provides downloadable datasets and interactive tools for exploration across studies (https://caninecancergenomeatlas.org/). CONCLUSION: By reprocessing heterogeneous studies through one analytical framework, the CCGA converts dispersed sequencing data into a reusable comparative resource. It distinguishes robust findings from observations limited by small cohorts and provides a practical foundation for studying cancer mechanisms in dogs, prioritizing hypotheses for veterinary precision oncology, and connecting naturally occurring canine cancers with human disease. Shared genomic alterations and direct comparisons between species support investigation of conserved oncogenic mechanisms, while downloadable data and interactive tools facilitate exploration and reuse. However, incomplete clinical annotation and uneven cohort representation limit clinical interpretation; and therapeutic hypotheses require further validation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



