Linking genotype, phenotype, and environment, beyond model organisms.
Most of what we know about how genes shape traits comes from a handful of model species. We extend those questions to wild plant lineages, from mangroves to milkvetches to sedges to cherries, by combining genomic, phenotypic, and environmental data drawn from field and herbarium specimens.
Our research focuses on understanding the processes that generate and maintain biodiversity in plants, with a focus on how species-specific phenotypic traits interact with the environment to influence the evolutionary processes driving spatial patterns of genetic variation. Specifically, we use bioinformatic and machine learning approaches to integrate genomic, environmental, morphological, geospatial, and digitized image data from field-collected and herbarium specimens. We computationally integrate these diverse data types to improve our understanding of the evolutionary processes that create and maintain biodiversity.
Study systems
We pursue these questions across a range of lineages and ecosystems:
Mangroves in Florida, the Caribbean, and Indonesia
Sedges of the Rocky Mountains
Canyon grape (Vitis arizonica), a wild grape in Southwest riparian corridors, where we the genomic and phenomic drivers that allow species to occupy immense elevational ranges
Milkvetches (Astragalus), where we use genomic data to test species boundaries
Prunus (Rosaceae), a genus of about 300 species including cherries, peaches, plums, and almonds, along with substantial understudied wild diversity
In Prunus, we use phylogenomic data from field and herbarium collections to resolve evolutionary relationships. We apply deep learning to detect morphological signatures of ancient hybridization and allopolyploidy, and to identify traits associated with tropical-to-temperate biogeographic transitions.
Research directions
An integrative framework for diversification. Developing analytical approaches that unite genomic, phenomic, and environmental data to identify the drivers of speciation.
The genomic basis of evolutionary innovation. Investigating the mechanisms behind major transitions, such as the shift from tropical to temperate climates and the repeated evolution of fleshy fruits from dry-fruited ancestors.
Prunus as a model clade. Establishing the cherry genus as a system for testing trait-based phylogeographic hypotheses, including how particular morphological characters shape adaptation to novel environments.
Connecting different data types inside museum specimens
Workflow for extracting multiple data types from specimens, ultimately resulting in functional traits that can be examined in phylogenetic contexts. Every specimen contains genomic data, which can be extracted from leaf tissue, environmental data, such as bioclimatic variables associated with the specimen’s georeferenced coordinates, and phenomic data, which can be measured using traditional morphometric approaches, or more recently, using computer vision-based machine learning algorithms that can extract phenotypic characters from images. Genomic data can be used to construct evolutionary trees, and phenotypic data and environmental data can be used to identify functional phenotypic traits. Here, traits such as entire versus serrated leaf margins (A), leaf shape and petiole length (B), and inflorescence architecture (C) can be advantageous in certain abiotic or biotic environments. Traits that are variable within extant species, or among extant sister species, can be used to tease apart the relative contributions of the genome and environment to shape these traits (denoted by asterisk). These key functional traits associated with recent divergences can be used to infer how phenotypic traits, the genome, and the environment interacted in more ancient divergences (denoted by hashtag).