Department of Genetics, Trinity College, University of Dublin, Dublin 2, IRELAND.
In this thesis I study how duplicate gene pairs created by a whole-genome duplication in an ancestor of several yeast species were resolved. I show that gene duplication may lead not just to the emergence of new gene functions, but also to the emergence of new species. I used comparative genomics between ten hemiascomycete yeasts to study both the process of gene loss that caused over 4000 genes to be rapidly lost from the S. cerevisiae genome and the altered molecular evolution of those genes that have been retained in duplicate. Among the genomes I studied was that of the non-model hemiascomycete yeast Kluyveromyces polysporus, which was sequenced, annotated and analyzed during the course of this thesis.
Three major findings arise from this work.
First, I show for the first time that both members of duplicate gene pairs experience a burst of protein sequence evolution in the immediate aftermath of duplication (Chapter 4). Following this burst, purifying selection is rapidly restored on one duplicate while the other continues to evolve rapidly for at least 100 Myr. Because gene duplication is often associated with the emergence of new biological functions, the altered evolutionary dynamics of duplicate genes identified in this work may be the molecular signature of evolutionary innovation.
Second, the work presented in Chapter 3 paints the most complete picture yet of gene loss in any organism. I show that when duplicate gene pairs are returned to single-copy the “choice” of which copy to lose is not random – as duplicate genes diverge in sequence, one member becomes favoured and will preferentially be retained, while the other is more likely to be lost. By contrast, for very young duplicate genes or those that are involved in highly conservative biological processes, selection cannot differentiate between the two copies and both are equally likely to be lost in independent lineages. The observation that natural selection can distinguish between copies of some duplicate gene pairs but not others suggests an analogy with the Nearly Neutral Theory, in which random genetic drift determines the fate of alleles whose selective coefficients are similar but natural selection is the dominant force when one allele confers a significant advantage over the other. A “nearly equal” theory of duplicate gene resolution may describe the process of gene loss after duplication.
Finally, I have provided the first evidence for a model of speciation in which ancestrally duplicated loci that have undergone reciprocal gene loss between a pair of species behave as Dobzhansky–Muller incompatibilities and contribute to reproductive isolation (Chapter 2). Because some spores produced after a hybridization between two lineages that have fixed null alleles at alternative copies of an ancestrally duplicated locus may inherit only these null gene copies and will be inviable assuming the gene is essential, gene duplication followed by gene loss can be a significant barrier to gene flow. Indeed, reciprocal gene loss gene loss at just 16 unlinked ancestrally duplicated loci is sufficient to reduce spore viability to ~1% but I show that hundreds of reciprocal gene losses separate all the major lineages that emerged after the WGD in yeast.