The abundance of orphan genes, or genes without known homologues, is amongst the greatest surprises uncovered by the sequencing of a large number of eukaryotic and bacterial genomes. It is therefore important to determine how the number of orphan genes will change as we sample new genomes. There are three possibilities. First, the number of orphans could continue to rise as we sample new genomes. Alternatively, orphan numbers could plateau in the future despite the sampling of novel taxa as has been suggested in the past (Siew & Fischer, 2003). Finally, the number could decrease by finding homes (gene families) for current orphans by improving our annotation methods and the sensitivity of our similarity searching algorithms (Skovgaard et al., 2001).

Here we examine these possibilities using data generated for a set of 122 bacterial species for which complete genomes are available. We have used these data to show that orphans are continuing to increase in number. This further emphasizes the importance of sequencing taxonomically diverse isolates (especially from environmental samples) to find novel predicted proteins. We suggest that orphans should now be classified as ‘taxonomically restricted genes’(TRGs), as this concept seems more useful for advancing our knowledge of these sequences and their potential ecological significance.