Fig. 3

Identification of CDS orthology groups: case of a gene whose transcripts are conserved over species. This figure concerns the gene presented in Figs. 1 and 2. Comparing its gene models over human, mouse and doc (see Fig. 1b) serves as a basis i) to compare the structures of CDSs across species (see Fig. 1a), ii) to predict orthologous CDSs using predicted exons and sites (see Fig. 1c), and iii) to establish orthology relations between CDSs (left of Fig. 3). Thus, the resulting graph of transcripts connects transcripts sharing the same CDS splicing structure, thereby identifying CDS orthology groups. Left of Fig. 3, the transcript graph for the orthologous genes ENSG00000001167,ENSMUSG00000023994 and ENSCAFG00000001580 contains 4 subgraphs, each being a triplet of orthologous CDSs (nodes are CDSs, blue for human, grey for mouse and red for dog, with “ ?” indicating a predicted CDS). This implies that i) all the seven known gene transcripts, in human, mouse and dog actually represent four different CDS splicing structures, ii) five new CDSs are predicted to make the graph complete, iii) for this gene, each of the four CDS splicing structures is feasible in human, mouse ang dog, so the three genes share the same orthologous CDSs. Right of Fig. 3, the four braces indicate details of the four CDS orthology groups, showing the Ensembl identifiers of known transcripts, the predicted CDSs, and the spliced CDS structures t1 t2 t3 and t4 shared over the three species