Genomics: More Than the Sum of the Parts

It has been known for some time that DNA composition varies across a given genome as well as between genomes (Filipski et al. 1973;Wagner and Capesius 1981). Genomic sequencing projects allow this observation to be confirmed at the sequence level (The Arabidopsis Genome Initiative 2000; Ashikawa 2001). However, the cause and function of these compositional differences are still obscure. Among the theories that may explain these phenomena (Eyre-Walker and Hurst 2001), mutation bias from C to T due to deamination of methylated C has been commonly used to account for them (Coulondre et al. 1978). As methylation is probably involved in a mechanism to silence transposable elements (Martienssen 1998), it makes sense that inactive methylated transposons can easily undergo C to T transition because they are under no selective pressure. However, this theory cannot explain other related compositional biases such as the CpG suppression observed in animal mitochondria (Cardon et al. 1994), where there is no DNA methylation.

The recent completion of the draft sequence of the genome of the riceindica subspecies (Yu et al. 2002) allowed Wong and coworkers (2002) to uncover a new kind of fine-scale GC heterogeneity. By analyzing in detail GC frequencies in a collection of rice full-length cDNAs and aligning them to the genome, they discovered that genes are richer in GC at the 5′ end than at the 3′ end. Interestingly, …