US-AUSTRALIAN ACADEMIES JOINT WORKSHOP

US-AUSTRALIAN ACADEMIES JOINT WORKSHOP ON VERTEBRATE COMPARATIVE GENOMICS
Beckman Conference Centre, Irvine, California, 23-25 May 2007

The evolutionary dynamics of mC in vertebrate genomes
by Gavin Huttley
Gavin Huttley is the head of the Computational Genomics laboratory in the John Curtin School of Medical Research at the Australian National University. His research seeks to apply evolutionary principles to biomedical problems. Specific areas of interest include: identifying factors affecting mutation risk; the evo-lutionary dynamics of epigenetic variation; the relationship between adaptive evolution and disease; and, development of statistical models for addressing these topics.

The modified base 5-methyl-cytosine (mC) plays a critical role in the vertebrate developmental program. Methylation of cytosine also greatly accelerates its de novo mutation rate to such an extent that a substantial proportion of all human single nucleotide polymorphisms may have arisen from mutation of mC. I will present work on the evolutionary dynamics affecting the characteristic sequence motif, CpG, at which mC occurs from both non-coding and protein coding vertebrate sequences. Two specific observations will be addressed: (i) CpG variability has been reported to vary across the human genome, increasing with sequence A+T%; (ii) the amino-acids encoded by CpG containing codons interchange in a modestly permissive manner. We use continuous time Markov models designed to distinguish potential confounding influences on estimation of substitutions. Our analyses of alignments of non-coding sequences from Human, Macaque, Dog, Mouse and Rat support the relationship between CpG variability and sequence composition. We show, however, that this relationship is shared by all nucleotides with a C/G neighbour, which is inconsistent with the previously proposed mechanisms. We speculate this effect can be accounted for by an influence of neighbouring nucleotides on DNA polymerase fidelity. Our analyses of alignments of protein coding genes from these species established that, for 46/64 examined genes, the substitution rate of nonsynonymous changes at CpG containing codons was significantly different from the average nonsynonymous rate. The rate parameter for CpG nonsynonymous changes was < 1 for all 64 loci, in contrast to that expected based solely from amino-acid analyses. The result suggests that despite the otherwise permissive interchange of the encoded amino-acids, a significant component of human polymorphism at CpG containing codons will be phenotypically influential.

Contact details:
Computational Genomics Group
The John Curtin School of Medical Research
GPO Box 334, Canberra City ACT 2601, Australia

Tel: +61 2 6125 7961
Email: Gavin.Huttley@anu.edu.au