Properties and distribution of pure GA-sequences of mammalian genomes

The article describes DNA sequences of mammalian genomes that are longer than 50 bases, but consist exclusively of G's and A's ('pure GA-sequences'). Although their frequency of incidence should be 10^-16 or smaller, the chromosomes of human, chimpanzee, dog, cat, rat, and mouse contained many tens of thousands of them ubiquitously located along the chromosomes with a species-dependent density, reaching sizes of up to 1300 [b]. With the exception of a small number of poly-A-, poly-G-, poly-GA-, and poly-GAAA-sequences (combined <0.5%), all pure GA-sequences of the mammals tested were unique individuals, contained several repeated short GA-containing motifs, and shared a common hexa-nucleotide spectrum. At most 2% of the human GA-sequences were transcribed into mRNAs; all others were not coding for proteins. Although this could have made them less subject to natural selection, they contained 160 times fewer point mutations than one should expect from the genome at large. As to the presence of other sequences with similarly restricted base contents, there were approximately as many pure TC-sequences as pure GA-sequences, but many fewer pure AC-, TA, and TG-sequences. There were practically no pure GC-sequences. The functions of pure GA-sequences are not known. Supported by a number of observations related to heat shock phenomena, the article speculates that they serve as genomic sign posts which may help guide polymerases and transcription factors to their proper targets, and/or as spatial linkers that help generate the 3-dimensional organization of chromatin.