SOX5P (Table 2). Other fragments corresponding to al., 2000) has been exploited in many laboratories to parts of SOX2,-17,-20, and-30 can be found in the identify approximately 30 vertebrate and over a dozen human genome, but these are short, lack an HMG box, invertebrate Sox genes or gene fragments. However, and contain gaps, insertions, or in-frame stop codons, the actual number of Sox genes in the mouse and human indicating that they are not segments of functional Sox genomes has remained unknown. With the availability genes (Table 2). No pseudogenes or pseudogene fragof complete drafts of these genome sequences, we can ments were found in the mouse genome. now determine the precise number of Sox genes, assign Sequence similarity between mouse Sox12 and hunames, and identify orthologs. This in turn provides a man SOX22 has been reported previously (Jay et al., basis for similar efforts in other model organisms as 1997; Bowles et al., 2000). The availability of the comsequence data become available. plete coding sequence reveals extensive non-HMG box In this analysis, we examined all published Sox sesequence homology between these two genes. This hoquences, and recent releases of the human and mouse mology, and the chromosomal location of both genes genome sequence from the relevant public sequencing within regions of conserved synteny, confirm that Sox12 consortia (Mouse Genome Assembly v3, 2 May 2002, and SOX22 are orthologs. Similar observations indicate http://www. ensembl. org and Human Genome Assembly that SOX20 and Sox15 (Bowles et al., 2000; Hiraoka et build 29, 5 April 2002, http://www. ncbi. nlm. nih. gov/ al., 1998) also are orthologs. We therefore rename hugenome/seq) and from Celera Genomics (Celera Discov- man SOX22 as SOX12 and human SOX20 as SOX15 ery System, indexed 2 May 2002, http://www. celera.(Table 1). com)(Lander et al., 2001; Venter et al., 2001). SOX pro- Our analysis suggests that no further nomenclature teins other than SRY were defined by the presence of changes or additions will be required for the mouse and the HMG domain signature sequence RPMNAFMVW human Sox family. The current system of nomenclature,(Bowles et al., 2000). Orthologous Sox genes were identi- loosely based on the order of gene discovery, is firmly fied by sequence similarity and chromosomal location entrenched in the literature, and the likely confusion within regions of conserved synteny, determined by and noncompliance associated with a more systematic nomenclature revision in our view outweigh the potential comparison of gene order. benefits. Our recommendations have been endorsed The mouse and human genomes were found to conby the HUGO Gene Nomenclature Committee (http://tain 20 orthologous pairs of Sox genes (Table 1). The www. gene. ucl. ac. uk/nomenclature). paired Sox genes show identical genomic organization,