Abstract
BACKGROUND: Functional diversification of genes in mammalian genomes is engendered by a number of processes, e.g., gene duplication and alternative splicing. Gene duplication is classically discussed as leading to neofunctionalization (generation of new functions), subfunctionalization (generation of a varied function), or pseudogenization (loss of the gene and its function). RESULTS: Here, we focus on the process of pseudogenization, but specifically for individual exons from genes. It is at present unclear to what extent pseudogenization of individual exon duplications affects gene evolution, i.e., is it a random phenomenon, or is it associated with specific types of genes and encoded proteins, and positions in gene structures? We gathered genomic evidence for pseudogenic exons (PsiEs, i.e., exons disabled by frameshifts and premature stop codons), to examine for significant trends in their distribution across four mammalian genomes (specifically human, cow, mouse and rat). Across these four genomes, we observed a consistent population of PsiEs, associated with 0.4-1.0% of genes. These PsiE populations exhibit codon substitution patterns that are typical of an endemic population of decaying sequences. In human, PsiEs have significant over-representation for functional categories related to 'ion binding' and 'nucleic-acid binding', compared to duplicated exons in general. Also, PsiEs tend to be associated with some protein domains that are abundant generally, e.g., Zinc-finger and immunoglobulin protein domains, but not others, e.g., EGF-like domains. Positionally, PsiEs are also significantly associated with the 5' end of genes, but despite this, individual stop codons are positioned so that there is significant avoidance of potential targeting to nonsense-mediated decay. In human, PsiEs are often associated with alternative splicing (in 22 out of 284 genes with PsiEs in their milieu), and can have different parts of their sequence differentially spliced in alternative transcripts. Some unusual cases of PsiEs embedded within 5' and 3' non-coding exons are observed. CONCLUSION: Our results indicate the types of genes that harbour PsiEs, and demonstrate that PsiEs have non-random distribution within gene structures. These PsiEs may function in gene regulation through generation of transcribed pseudogenes, or regulatory alternate transcripts.