How significant is 98.5% 'junk' in mammalian genomes?

Conversion of genetic information from RNA into DNA by reverse transcription is ancient and was instrumental for the transition from the RNP world to modern cells (Brosius, 1999a, 2003). Surprisingly, in many eukaryotic lineages, the process of retroposition is still very active. All types of RNAs (Brosius, 1999b) can be reverse transcribed and their cDNA copies reintegrated into genomes as retronuons (a nuon is any discrete segment of nucleic acid (Brosius and Gould, 1992)). About 38% or 42% of the mouse and human genomes, respectively, consist of discernible retronuons (excluding mRNA-derived retroposons). Only 1.5% of the human genome consists of exons coding for proteins. Even considering DNA transposons (∼1–3% discernible), slippage during replication and a relatively large fraction derived from segmental duplications it is conceivable that the remainder of mammalian genomes is probably derived from ancient, today non-discernible retronuons. Hence, the vast majority of mammalian genomes have been contributed by retroposition. Retroposition predominantly leads to ‘junk DNA’. However, mRNA-derived retronuons are known to give rise to active genes, often with different expression patterns than their respective founder genes (Brosius and Gould, 1992; Brosius, 1991). Retronuons derived from small non-messenger RNAs (snmRNAs) generate novel snmRNA genes (such as the neuron-specific BC1 and BC200 RNAs) (Brosius, 1999b; Brosius and Gould, 1992). Frequently, retronuons are exapted (co-opted) as regulatory elements that may alter expression or processing of targeted genes (Brosius and Gould, 1992) (for compilations see http://www-ifi.uni-muenster.de/exapted-retrogenes/tables. html). Consequently, retronuons are a major driving force of evolution and perhaps even speciation. Comparison of the human genome with that of other mammals such as mouse or, in particular, chimpanzee reveals that neither contains numerous additional genes. Instead, one observes exaptation of novel exons (often involving alternative splicing) from previously nonaptive intronic (as predicted by Gilbert, 1978) or flanking sequences originally generated by retroposition. Furthermore, the differential expression of shared genes with respect to developmental onset and/or cell-type specificity, that is triggered by de novo insertions of retronuons, will turn out to be a recurrent theme in species differences at the genomic level.