Figure 4.

Manual annotation of selected orthologous loci with human-mouse gene structure distinctions. Positive-strand transcription, relative to the genome assembly, is in red. Negative-strand transcription, relative to the genome assembly, is in blue. Beige boxes delineate bidirectional promoters (BDP) and sense-antisense overlaps (SAS). A 5′/5′ SAS is an overlap of two genes at their 5′ ends (a divergent overlap). A 3′/3′ SAS is an overlap of two genes at their 3′ ends (a convergent overlap). (A) Two protein-coding genes have orthologs: PITX and H2AFY. H2AFY has a positionally equivalent [see Babak et al. (2007) for definition] SAS lncRNA at its 3′end (AK026965 in human) in both species, suggesting a sequence-independent requirement for SAS pairing of H2AFY. In human, H2AFY shares a bidirectional promoter with another lncRNA (AK092789) for which no genomic or transcriptional conservation exists in mouse. (Supplementary Dataset 5: rows 62–63.). (B) The human protein-coding gene AY358799 has a mouse ortholog, “Ncrna00085” (encoding a 339-aa protein, despite its misleading name that arose out of incorrect public “lincRNA” annotations that are loaded into the UCSC Genome Database). The same cluster of three conserved microRNAs is observed immediately upstream of this gene in both species. However, this protein-coding gene has a SAS lncRNA, AK125996, only in human. Despite the more comprehensive mouse cDNA/EST coverage by the FANTOM3 data, no antisense cDNAs or ESTs are found at the orthologous mouse locus. (Supplementary Dataset 5: rows 14–15.). (C) The human TSSC4 gene overlaps a SAS lncRNA, AK095568, at its 5′ end. The human TSSC4 and TRPM5 genes are clearly separated along the genome, with no intervening transcription. Mouse Tssc4 is SAS to an extended 3′-end isoform of Trpm5, and also lacks any cDNA or EST evidence of a 5′-end SAS transcript. The nearby CD81 gene has a conserved SAS lncRNA in human and mouse. (Supplementary Dataset 5: rows 68–69.).