Adventures in Aloha
Role of gene flow in radiations of fish
Adaptive radiations are central to our understanding of evolution because they generate a wealth of ecological, phenotypic, and species diversity in rapid bursts. The availability of resources in new environments with few competitors has long been seen as the major force driving adaptive radiations, but it is a longstanding question why only some lineages rapidly diversify in response to such ecological opportunities while others do not. A growing number of studies have identified extensive hybridization and introgression across a range of adaptive radiations, contributing to the emerging view that gene flow may play a key role in adaptive radiation. However, it is still unclear how often hybridization is necessary for rapid diversification, as opposed to simply being pervasive throughout the history of any young rapidly diversifying group. My research interests lie in examining the role this pervasive hybridization has had on diversification processes in radiations of fish and whether it potentially triggered those radiations using population genomic approaches and experimental hybridization.
Below are some ramblings about the work I have done so far in addressing the contributions of gene flow into two different radiations of fish.
San Salvador Island pupfish
Pupfish species of the genus Cyprinodon inhabit saline lakes and coastal areas across the Caribbean and Atlantic and nearly all pupfishes are allopatric, dietary generalists consuming algae and small invertebrates. In contrast, the hypersaline lakes of San Salvador Island, Bahamas house a small radiation of Cyprinodon species, including two endemic specialists. These specialists have adapted to unique trophic niches using novel morphologies: the molluscivore Cyprinodon brontotheroides with a unique nasal protrusion and the scale-eater Cyprinodon desquamator with enlarged oral jaws. (There may even be a fourth species on the island of 'small-jawed scale-eaters', but stay tuned for more on them in future research projects).
Crescent Pond, San Salvador Island Bahamas: home to two specialist pupfish
Beneath the surface of Crescent Pond: shallow, hot, lots of green and a bit of mud. Perfect for pupfish
Generalist pupfish Cyprinodon variegatus
Snail-eating specialist pupfish Cyprinodon brontotheroides
Scale-eating specialist pupfish Cyprinodon desquamator
The currently available ecological and genetic data on the group provides little indication as to why this radiation is localized to a single island.Variation in ecological opportunity among hypersaline lake environments in the Caribbean does not appear to explain the rarity of this radiation. This finding suggests a potentially important role for sufficient genetic variation to respond to abundant, underutilized resources in these environments.
In our investigations of the role of gene flow in this radiation so far, we found evidence of extensive and previously unexpected variation in evolutionary relatedness among Caribbean pupfish, some of which appears to be driven by adaptive introgression events in the history of this radiation (Richards and Martin 2017).
Relationship we expect to see across pupfish genomes: San Salvador Island species are more closely related to each other than other pupfish species. This particular tree was assigned to 64% of the genome.
Unexpected relationships also found across the genome though! These examples represent places in the genome where one of the specialist on San Salvador is more closely related to outgroup species than to the other San Salvador pupfish. May be caused by introgression with outgroups!
We also find evidence of introgression using formal test statistics for introgression. While adaptive introgression appears to be rare across the genomes of the San Salvador species, it may have introduced adaptive variants important in the evolution of the complex phenotypes of the specialists. Four of the top 11 candidate adaptive introgression regions contain genes with known effects on jaw morphology in zebrafish or associated with pupfish jaw size, the primary axis of phenotypic divergence between species in this system (Richards and Martin 2017). Our findings that multiple sources of genetic variation contribute to the San Salvador radiation suggests that a complex suite of factors, including hybridization with other species, may be necessary for adaptive radiation in addition to ecological opportunity.
One of the candidate introgressed regions, where we see strong signatures of introgression (fd statistic), divergence between San Salvador Island specialists (Fst, Dxy), and selective sweeps (pi and Tajima's D). This particular region features the gene ski, which has strong craniofacial effects in zebrafish and mice, some of which are uncannily similar to the facial structure we see in the molluscivore specialist!
Coming down the production line: 200+ whole genomes of pupfish species from across the Caribbean for geographic distributional analyses of genetic variation in Caribbean, demographic analyses of timing of gene flow events in San Salvador, and experimental hybridization between Caribbean lineages to look for potential transgressive segregation due to hybridization resulting in novel phenotypes and/or genetic incompatibilities that may have contributed to the San Salvador radiation. Stay tuned for more!
Cameroon crater lake cichlids
The crater lake cichlids of Cameroon are some of the most convincing examples of sympatric speciation we have in nature, based on classic criteria of sympatric ranges, monophyly (suggesting a single colonization of the lake), reproductive isolation, and little chance of allopatric periods given the uniform shapes of the crater lake and isolation from riverine sources. However, genomic data has recently provided evidence of complex histories of multiple colonizations of cichlids in these crater lakes than previously thought. This is a growing trend. Most sympatric speciation case studies revisited with genomic data show similarly complicated histories of colonization, casting doubt on all of our examples of sympatric speciation in the wild. However, much of this evidence comes from genetic data taken 'randomly' from across the genome and are based on statistics that represent genome-wide averages. This genome-wide evidence can tell us that there was multiple colonizations of the lake and gene flow between populations that experienced a period of allopatric separation, but not how it influence speciation within the crater lake. It is possible that this gene flow initiated speciation, aided stalled speciation events to completion, or was neutral to speciation.
I am currently investigating the role secondary gene flow played with regards to the cichlid radiation in Barombi Mbo, an ecological diverse 11 species radiation. This radiation includes cichlids with some interesting specializations, like deep water specialist Konia dikume and sponge-eating specialist Pungu maclareni. Using whole genomes from Barombi Mbo and nearby riverine cichlids, I investigated where in the genome these signals of secondary gene flow from riverine populations are coming from. Based on functional annotations of these regions, there is little evidence that secondary gene flow from riverine populations brought in genetic variation contributing to traits like jaw morphology, body pigmentation, retinal pigmentation (Richards, Poelstra & Martin 2018). So our current evidence leans towards not completely ruling out sympatric speciation in Barombi Mbo just yet. However, there are other ways introgressed variation could have contributed to speciation (regulatory elements, genetic incompatibilities, etc), and we are working on investigating other genomic lines of evidence for sympatric speciation (also the subject of an opinion piece we are writing; Richards, Servedio & Martin 2018).
Heading 1
Relationship we expect to see across Barombi Mbo cichlid genomes (black lineages) if they fit in classic criteria for sympatric speciation (monophyly/single colonization of the lake). This particular tree was assigned to 50% of the genome. Colored lineages represent closely related riverine and lake populations of Sarotherodon.
Phylogenetic evidence of non-monophyletic relationships among Barombi Mbo species and riverine outgroups, where species in Barombi Mbo appear more closely related to riverine outgroups than to other Barombi Mbo species. These sort of relationships are suggestive of multiple colonizations of the lake over time, casting doubt on sympatric speciation based on the classic criteria. However, these relationships only represent a very small proportion of the genome (less than 1% of it), and closer investigation of genes in those regions doesn't reveal any obvious role for introgressed material contributing to speciation within the lake.