Fig. 2
Genomics analysis of canine GWA mapped loci. All mapped intervals and up to 250 kb on each side were considered for genomics evidence of osteosarcoma, osteoblast or cancer genes. Only candidates with at least one hit are shown
1All loci were mapped by Karlsson et al. 2013 (PMID: 24330828) except one (FGF9 in this study); the BICF2S23637753 locus was mapped in this work using the same published data. In parentheses, the first breed is the discovery GWA breed (Karlsson et al. 2013, PMID: 24330828); the second or third are breeds fixed for the risk allele (derived in this work from Karlsson et al. 2013, PMID: 24330828, except Chr11:41 reported in that study).
2All genes with at least one hit for biological relevance categories here are shown. Genes most implicated by biological relevance and modeling are bold; genes that are not positionally top candidates at a locus are italicized. *JCAD is the official HUGO gene symbol, but most genomic data refers to it as KIAA1462; **refers to a miRNA cluster important in human cancer but is not represented in the biological relevance data types considered here; ***refers to data for the Ewing sarcoma gene EWSR1, the parent gene for canid-specific retrogene EWSR1CR; ****at 700 kb away, TUSC3 is the only gene beyond 200 kb outside the published interval, but is included here because it is an candidate OS tumor suppressor in dogs (PMID: 21837709).
3Transposon-based forward genetic screen for osteosarcoma development and metastasis (413 genes; Moriarity et al. 2015, PMID: 25961939)
4mRNA-sequening based osteosarcoma expression studies in humans, mice and dogs (Scott et al. 2018, PMID: 29066513)
5Cancer Index is a curated dataset of cancer types and associated germ line or somatic variant genes based on diverse types of evidence (contains 2168 cancer genes and 48 OS genes; Cotterill S.J. 2015, Cancer Genetics Web: http://www.cancer-genetics.org/)
6Transcriptomics of aggressive osteosarcomas in humans, dogs and mice (used genes with p < 0.05, FDR =/< 0.05: 3500 murine and 492 canine; Davis et al. 2017, PMID: 29100308)
7Human recurrent somatic copy number alterations in osteosarcoma (only observed overlap was in deleted genes) (Perry et al. 2014, PMID: 25512523)
8The integrated encyclopedia of DNA elements in the human genome, ENCODE Project Consortium, 2012 (PMID: 22955616)
9Single-cell RNA-sequencing analysis of the murine growth plate for bone elongation and regeneration (9739 genes; Li et al. 2016; PMID: 27160914)
10Statistical analysis of > 8200 human tumor-normal pairs in diverse cancer types to identify likely cancer driver properties (used top 1524 oncogenes, 1071 tumor suppressors; Davoli et al. 2013, PMID: 24183448)
11The Candidate Cancer Gene Database is a database of cancer driver genes from forward genetic screens in mice (9485 genes; Abbott et al. 2015, PMID: 25190456)
12Cancer Gene Census is a very stringent catalog of genes which contain mutations that have been causally implicated in cancer (719 genes; Futreal et al. 2004, PMID: 14993899)
13The number of hits for biological relevance categories marked by black boxes, but not heterochromatin (in grey)
14Number of hits searching each gene in the full Gene Expression Omnibus (Edgar et al. 2002, PMID: 11752295) for that term as a gene symbol together with the term osteosarcoma
15Modeling data from this work were arbitrarily classified into tier 1 (Odds Risk, OR > 9), 2 (OR 3–7.5) and 3 (OR 1.1–2.9)
16Comparative Genomic Hybridization (CGH) copy number alteration analysis by Thomas and Breen (Karlsson et al. 2013, PMID: 24330828)
17Allele frequencies of fixed or nearly-fixed risk and non-risk peak alleles were derived in this work using the Karlsson et al. 2013 genotype data (PMID: 24330828)