Biology subjects

Harrison, J. G.

Publications and source records attributed to Harrison, J. G..

6 recordsLinked to original sources

bioRxiv2024

Interpretable and predictive models to harness the life science data revolution

The proliferation of high-dimensional data in ecology and evolutionary biology raises the promise of statistical and machine learning models that are highly predictive and interpretable. However, high-dimensional data are commonly burdened with an inherent trade-off: in-sample prediction of outcomes will improve as additional variables are included in the model, but this may come at the cost of poor predictive accuracy and limited generalizability for future or unsampled observations (out-of-sample prediction). To confront this problem of overfitting, sparse models can focus on key variables by correctly placing low weight on unimportant variables. We competed nine methods to quantify their performance in variable selection and prediction using simulated data with different sample sizes, numbers of variables, and strengths of effects. Overfitting was typical for many methods and simulation scenarios. Despite this, in-sample and out-of-sample prediction converged on the true predictive target for simulations with more observations, larger causal effects, and fewer variables. Accurate variable selection to support process-based understanding will be unattainable for many realistic sampling schemes in ecology and evolution. We use our analyses to characterize data attributes for which statistical learning is possible, and illustrate how some sparse methods can achieve predictive accuracy while mitigating and learning the extent of overfitting.

genomics↗

bioRxiv2022

The abundances and occurrences of foliar microbes are poorly predicted by variation in plant traits and abiotic conditions

Much effort has been made to understand why foliar microbes live where they do. However, whether foliar microbiome composition can be predicted is unknown. Here, we determine the limits of prediction using metabarcoding data of both fungal and bacterial assemblages that occur within (endophytes) and without (epiphytes) leaves from 59 plant taxa. We built random forest models for prevalent taxa and quantified the combined predictive power of 24 plant traits, 12 abiotic conditions and 7 additional features. As response variables, we considered microbial relative and absolute abundances, and occurrences. Most microbial taxa were too rare to effectively model, but model performance was generally poor even for the most prevalent and abundant taxa (model R2 was typically <0.1). Fungi were more tractable for modeling than bacteria. Models of Shannons diversity were moderately successful but those for richness were not. Taxa responded idiosyncratically and non-linearly to variation in the foliar habitat. When prevalent microbes were included as features in models, performance improved. Our results suggest that easily measurable aspects of the phyllosphere habitat are poor predictors of microbiome composition. These results pose a challenge for the study of microbial biogeography and we discuss possible ways forward.

ecology↗

bioRxiv2022

Additive genetic effects in interacting species jointly determine the outcome of caterpillar herbivory

Plant-insect interactions are common and important in basic and applied biology. Trait and genetic variation can affect the outcome and evolution of these interactions, but the relative contributions of plant and insect genetic variation and how these interact remain unclear and are rarely subject to assessment in the same experimental context. Here we address this knowledge gap using a recent host range expansion onto alfalfa by the Melissa blue butterfly. Common garden rearing experiments and genomic data show that caterpillar performance depends on plant and insect genetic variation, with insect genetics contributing to performance earlier in development and plant genetics later. Our models of performance based on caterpillar genetics retained predictive power when applied to a second common garden. Much of the plant genetic effect could be explained by heritable variation in plant phytochemicals, especially saponins, peptides, and phosphatidyl cholines, providing a mechanistic understanding of variation in the species interaction. We find evidence of polygenic, mostly additive effects within and between species, with consistent effects of plant genotype on growth and development across multiple butterfly species. Our results inform theories of plant-insect coevolution and the evolution of diet breadth in herbivorous insects and other host-specific parasites. Teaser summaryThe combined, additive effects of plant and insect genetic variation explain Melissa blue caterpillar growth and development on alfalfa plants.

genomics↗

bioRxiv2019

Whole-genome duplication and host genotype affect rhizosphere microbial communities in Arabidopsis thaliana

The composition of complex microbial communities found in association with plants is influenced in part by host phenotype. Yet, the salient genetic architecture is often unknown. Genome duplication events are common in the evolutionary history of plants, influence many important plant traits, and may affect associated microbial communities. Using experimentally induced whole genome duplication (WGD), we tested the effect of WGD on rhizosphere bacterial communities in Arabidopsis thaliana. Specifically, we performed 16S rRNA amplicon sequencing to characterize differences between microbiomes associated with specific host genotypes (Columbia vs. Landsberg) and ploidy levels (diploid vs. tetraploid). We modeled abundances of individual bacterial taxa by utilizing a hierarchical Bayesian framework, based on the Dirichlet and multinomial distributions. We found that host genotype and host ploidy level affected rhizosphere community composition, for instance, the microbiome of the tetraploid Columbia genotype differed from that of other host genotypes. We then tested to what extent microbiomes derived from a given host genotype or ploidy level affected plant performance by inoculating sterile seedlings of each genotype with microbial communities harvested from a prior generation. We found a negative effect of the tetraploid Columbia microbiome on growth of all four plant genotypes. The findings suggest that while both host genotype and ploidy affect microbial community assembly, bacterial communities found in association with only some host genotypes may affect growth of subsequent plant generations. ImportancePlants influence the composition of their associated microbial communities; yet the underlying host genetic factors are often unknown. Genome duplication events are common in the evolutionary history of plants and affect many plant traits, including the quality and quantity of compounds exuded into the root zone, which can affect root-bound microbes. In Arabidopsis thaliana, we characterized how whole-genome duplication affected the composition of rhizosphere bacterial communities, and how bacterial communities associated with two host plant genotypes and ploidy levels affected subsequent plant growth. We observed an interaction in which ploidy level within one host genotype affected both bacterial community composition and function. This research reveals how genome duplication, a widespread genetic feature of both wild and crop plant species, influences the coexistence of bacterial taxa and affects plant growth.

plant biology↗

bioRxiv2019

The biogeography and biodiversity of endophytes--how far have we come and where do we go from here?

The interiors of plants are colonized by diverse microorganisms that are referred to as endophytes. Endophytes have received much attention over the past few decades, yet many questions remain unanswered regarding patterns in their biodiversity at local to global scales. To characterize research effort to date, we synthesized results from [~]600 published studies. Our survey revealed a global research interest and highlighted several gaps in knowledge. For instance, of the seventeen biomes encompassed by our survey, seven were understudied and together composed only 7% of the studies that we considered. We found that fungal endophyte diversity has been characterized in at least one host from 30% of embryophyte families, while bacterial endophytes have been surveyed in hosts from only 10.5% of families. We complimented our survey with a vote counting procedure to determine endophyte richness patterns among plant tissue types. We found that variation in endophyte assemblages in above-ground tissues varied with host growth habit. Stems were the richest tissue in woody plants, whereas roots were the richest tissue in graminoids. For forbs, we found no consistent differences in relative tissue richness among studies. We propose future directions to fill the gaps in knowledge we uncovered and inspire further research. Originality-Significance StatementMuch remains to be learned regarding the biodiversity and distribution of the microbes that colonize the interiors of plants. Here, we surveyed approximately 600 publications to characterize gaps in knowledge and provide a roadmap for future research. We compared biomes, plant families, and geographical regions in terms of the research interest that they have garnered. Additionally, we synthesized published results and report that variation in endophyte richness among plant tissue types is a function of host growth habit. Stems were the richest tissue in woody plants, whereas roots were the richest tissue in graminoids. We hope to inspire research to fill the gaps in knowledge that we uncovered.

ecology↗

bioRxiv2018

Caterpillars on a phytochemical landscape

Modern metabolomic approaches that generate more comprehensive phytochemical profiles than were previously available are providing new opportunities for understanding plant-animal interactions. Specifically, we can characterize the phytochemical landscape by asking how a larger number of individual compounds affect herbivores and how compounds covary among plants. Here we use the recent colonization of alfalfa (Medicago sativa) by the Melissa blue butterfly (Lycaeides melissa) to quantify plant metabolites and the performance of caterpillars as affected by both individual compounds and suites of covarying phytochemicals. We find that survival, development time and adult weight are all associated with variation in nutrition and toxicity, including biomolecules associated with plant cell function as well as putative anti-herbivore action. The plant-insect interface is complex, with clusters of covarying compounds in many cases encompassing divergent effects on different aspects of caterpillar performance. Individual compounds with the strongest associations are largely specialized metabolites, including alkaloids, phenolic glycosides and saponins. The saponins are represented in our data by more than 25 individual compounds with beneficial and detrimental effects on L. melissa caterpillars, which highlights the value of metabolomic data as opposed to approaches that rely on total concentrations within broad defensive classes.

ecology↗