Molecular convergence at the sequence level: a genome-wide approach in a novel mammalian model

Convergent evolution is the independent origin of the same feature in different groups of living things. Classic examples include the vertebrate wing, which has independently evolved a number of times, for example in bats, birds and pterosaurs, and the similar image-forming eyes of vertebrates and some invertebrates such as squid. The fact that similar structures have evolved several times suggests that they evolved to perform similar functions, so convergent evolution is powerful evidence that natural selection has shaped these features - there can be little doubt that bat and bird wings both evolved to allow powered flight, for example. Yet though examples of convergence are extremely common in the tree of life, we understand very little about the extent to which convergent evolution happens at the genetic level, in sequences of DNA and the proteins that they code for. We have recently identified several examples of apparent convergence in a suite of genes involved in hearing in different groups of echolocating mammal. Echolocation involves the production of sonar pulses and processing of the returning echoes for hunting and orientation, and poses particular challenges for high frequency hearing. It is seen at its most sophisticated in some lineages of bats and whales. For many years, echolocating bats were separated from fruit bats; however, advances in our ability to resolve species' relationships provided irrefutable evidence that some echolocating bats were in fact more related to the fruit bats than they were to each other. This finding has led to a revision of bat evolutionary relationships; so raising the issue that echolocation has either been lost by the fruit bats, or has evolved more than once by convergence. We have studied 'hearing genes' in bats and whales, and found that evolutionary trees based on four of these genes all unite echolocating bats into a single but technically incorrect group. Even more surprising, one of these genes leads to a well supported group of these bats with echolocating dolphins. These results raise the intriguing possibility that convergence in anatomical traits might sometimes be underpinned by convergence at the sequence level. Any finding of convergence of this kind is surprising, as the number of possible sequences for any gene is astronomically large. Therefore, such cases are unlikely to arise by chance. The identification of convergent molecular evolution in a number of different genes associated with a particular trait is, to our knowledge, unprecedented. Our evidence suggests that molecular convergence may be far more common than currently suspected. This may be partly because few scientists have been looking for this kind of convergence, and there has been no systematic attempt to investigate how common it might be. Methods to detect convergence in DNA or protein sequences are also relatively new. To confirm our finding, we want to search for convergent sequences across entire genomes, looking for genes that show convergence between the groups of echolocating bats, between bats that share similar echolocation calls, and also between bats and whales. If convergence is common in this system, it will advance our knowledge of echolocation, for example by identifying a number of genes probably involved in this system. More importantly, if confirmed in other systems, it will change the way scientists think about how genes and proteins evolve, suggesting that the pathways that evolution can take may be more constrained than previously thought, so that there may be relatively few good ways for evolution to fashion a protein to do a particular job.

Cases of adaptive functional and structural convergence, where different lineages evolve similar traits independently by natural selection, have proven central to our understanding of evolution. Yet convincing reports of molecular convergence at the sequence level are exceptionally rare. This paucity of cases might either be due to a genuine lack of sequence convergence, or could simply reflect under-reporting because of our inability to detect and test for this phenomenon. The independent evolution of echolocation in some bat lineages, and in toothed whales, represents a spectacular example of multiple convergence in mammals that has led to numerous shared auditory features among unrelated taxa. We have unpublished data from four different 'hearing genes' that show sequence convergence among unrelated groups of echolocators. All four gene trees unite paraphyletic echolocating bats into one clade, and one unites echolocating bats and dolphins. Using echolocation as a model system, we will develop a novel pipeline to test for convergence at the sequence level at multiple taxonomic levels. Our approach will utilise deep sequencing technology and phylogenetics to assess individual site-wise support (nucleotides and amino acids) for competing true versus convergent phylogenetic hypotheses. Our pilot data suggest that the pathways that evolution can take may be more constrained than previously thought, and we anticipate our results from genome-wide scans for convergence will change the way scientists think about how genes and proteins evolve. Our project will have benefits for those working in genetics, bioinformatics and comparative genomics, and have potential applications for the detection of disease-causing mutations.

Who will benefit? Our proposal tackles a pivotal question in evolution: to what extent has the evolution of convergent phenotypic traits arisen by common genetic routes? Our outputs (incl. a pipeline developed for its diagnosis) will be of interest and benefit to multiple user-groups include phylogeneticists, evolutionary geneticists, genome biologists, bioinformaticians, as well as researchers in echolocation, hearing and bats/cetaceans. Indeed, an ability to resolve true species relationships is vital in forming a comparative framework against which much biological data must be understood. Our 2X genome data for 4 new bat spp. will ensure that as well as addressing our specific aims, our invested time and money will have enormous impacts (publicly archived genetic data are used by 1000s of researchers daily). Medical geneticists will also derive indirect benefits. Comparative genetic data frequently inform our understanding of human disease, and vice versa. Indeed, genes implicated in diseases/developmental abnormalities in humans often show mutations in other animal models with similar conditions, and sometimes have undergone evolutionary changes in other lineages (e.g. opsin genes with blindness-causing mutations in humans have been turned off in some nocturnal taxa). There is also now interest in diagnosing whether cancers and other diseases arise via parallel/convergent or different somatic mutations across individuals. Bat genome data are arguably especially interesting given the evolutionary innovations seen in bats, and their potential implications for understanding processes in other taxa (e.g. wing development/limb deformities; echolocation/deafness; hibernation/fat metabolism). How will they benefit? Our findings will have immediate and major impacts in the fields of evolution, phylogenetics and zoology. As well as addressing a major gap in current knowledge, our methodological pipeline for testing for sequence convergence will also advance current capacity to process genomic data, so contributing to the basic toolkit of comparative biology. In the longer-term, the publicly accessible genetic and genomic data generated will provide immeasurable benefit for the users listed. Indeed, the impact of these data will extend well beyond our personal research interests. Updated annotations will ensure our generated data provide a valuable evolving resource for scientists worldwide, with the potential to enhance knowledge of gene function and evolution. Our PDRA will be trained in handling and assembling data from Next Generation Sequencing, as well as downstream analyses. These competitive skills are already lacking in the UK workforce, and are likely to become critical as new genome data are generated at ever increasing rates. Our PDRA's skills will benefit our national research base, and will be transferable to a wide range of sectors. What will be done to ensure they benefit? To ensure other researchers derive maximum benefit from our findings and data, we will make all parts of our results freely available. Briefly, raw sequence data will be uploaded to GenBank, following protocol. All source code for our pipeline for diagnosing convergent sequence evolution, and our tree files, will be hosted on our server for free download. Cotton already hosts source code on his web page. Our processed data will be written up and submitted to high impact journals. Where possible, we will pay for Open Access publication. We will also present our findings at national and international meetings. To communicate results that are of wider or public interest, we will work with the press offices (QMUL, BBSRC and journals publishing our work). We have track records in writing press releases, with our work covered by national media (newspapers/radio e.g. BBC), journal editorials or reviews (e.g. Nature, Current Biology, TREE) and internal publications (e.g. NERC) (see Rossiter's web site).