Thursday 21 February 2019, 1.00PM
Speaker(s): Professor Laurence D. Hurst, University of Bath
Is the human genome a perfect and well functioning machine or is it sloppy and error prone? This question is key to understanding both how our genome works and to understanding which mutations might or might not cause disease.
In this talk I will consider this issue by considering how our genes evolve. We often presume, for example, that missense mutations are commonly under selection and that this is because they swap one amino acid for another. Conversely, there is long tradition that assumes that exonic synonymous mutations cannot be deleterious and cannot cause disease because they don’t affect the protein. I’ll illustrate that both arguments are likely to be substantially incorrect. Importantly, our gene sequences specify not simply the order of amino acids but, at the DNA and RNA level, specify other motifs. I’ll concentrate on the impact of exonic splice enhancer motifs (ESE), key to the control of splicing. These motifs are at a high density, 30-40% of sequence at exon ends, highest in proximity to large introns.
As the human genome is unusual in having many small exons and large introns, our ESE density is unusually high. ESEs skew amino acid and codon usage and affect rates of evolution, reducing the rate of synonymous evolution by as much as 15-30%, predominantly at exon ends. As a predictor of the rate of protein evolution, the proportion of sequence near exons ends is as strong a predictor as the best universal predictor of protein evolutionary rates (expression breadth). Many mutations that might otherwise be considered innocuous thus can have a major phenotypic effect. From the ClinVar dataset we estimate that 25-45% of all disease-associated mutations have their effects via splicing modification and show that disease-associated mutations are most common in “fragile” exons, those with low ESE density and hence low resilience to mutation.