Nonrandom distribution of Alu elements in genes of various functional categories: insight from analysis of human chromosomes 21 and 22

The first draft of the human genome has revealed enormous variability in the global distribution of Alu repeat elements. There are regions such as the four homeobox gene clusters, which are nearly devoid of these repeats that contrast with repeat dense regions in other transcriptionally active regions of the genome. Our analysis of the completely sequenced chromosomes 21 and 22 revealed a striking bias in Alu distribution. These elements are more clustered in genes which are involved in metabolism, transport, and signaling processes. In contrast, they are significantly fewer in genes coding for information pathway components as well as structural proteins. This bias in Alu distribution is independent of the effect of Alu density of the flanking genomic region and is also not affected by the GC content of the gene and its upstream and downstream regions. The relative proportions of Alu subfamilies (Alu J, Alu S, and Alu Y) are not significantly different in genes with high Alu density belonging to the functional categories of transport, metabolism, and signaling. However, in the structural proteins and information genes, these proportions are lower than the other three categories. We suggest that Alu elements might be involved in regulatory mechanisms and are therefore differentially selected in primate genomes