Michael Broe

Tissue-Specific Venom Composition and Differential Gene Expression in Sea Anemones

Cnidarians represent one of the few groups of venomous animals that lack a centralized venom transmission system. Instead, they are equipped with stinging capsules collectively known as nematocysts. Nematocysts vary in abundance and type across different tissues; however, the venom composition in most species remains unknown. Depending on the tissue type, the venom composition in sea anemones may be vital for predation, defense, or digestion. Using a tissue-specific RNA-seq approach, we characterize the venom assemblage in the tentacles, mesenterial filaments, and column for three species of sea anemone (Anemonia sulcata, Heteractis crispa, and Megalactis griffithsi). These taxa vary with regard to inferred venom potency, symbiont abundance, and nematocyst diversity. We show that there is significant variation in abundance of toxin-like genes across tissues and species. Although the cumulative toxin abundance for the column was consistently the lowest, contributions to the overall toxin assemblage varied considerably among tissues for different toxin types. Our gene ontology (GO) analyses also show sharp contrasts between conserved GO groups emerging from whole transcriptome analysis and tissue-specific expression among GO groups in our differential expression analysis. This study provides a framework for future characterization of tissue-specific venom and other functionally important genes in this lineage of simple bodied animals.

Multi-copy venom genes hidden in de novo transcriptome assemblies, a cautionary tale with the snakelocks sea anemone Anemonia sulcata (Pennant, 1977)

Using a partial transcriptome of the snakelocks anemone (Anemonia sulcata) we identify toxin gene candidates that were incorrectly assembled into several Trinity components. Our approach recovers hidden diversity found within some toxin gene families that would otherwise go undetected when using Trinity, a widely used program for venom-focused transcriptome reconstructions. Unidentified hidden transcripts may significantly impact conclusions made regarding venom composition (or other multi-copy conserved genes) when using Trinity or other de novo assembly programs.

An Introduction to Linear Mixed Models

This chapter introduces linear mixed models at an elementary level. The prerequisite for understanding this presentation is a basic knowledge of the foundational ideas of linear regression discussed in the preceding chapter.

Microbiome Composition and Diversity of the Ice-Dwelling Sea Anemone, Edwardsiella andrillae.

Edwardsiella andrillae is a sea anemone (Cnidaria: Anthozoa: Actiniaria) only known to live embedded in the ice at the seawater interface on the underside of the Ross Ice Shelf, Antarctica. Although the anatomy and morphological characteristics of E. andrillae have been described, the adaptations of this species to the under-ice ecosystem have yet to be examined. One feature that may be important to the physiology and ecology of E. andrillae is its microbiome, which may play a role in health and survival, as has been deduced in other metazoans, including anthozoans. Here we describe the microbiome of five specimens of E. andrillae, compare the diversity we recovered to that known for temperate anemones and another Antarctic cnidarian, and consider the phylogenetic and functional implications of microbial diversity for these animals. The E. andrillae microbiome was relatively low in diversity, with seven phyla detected, yet included substantial phylogenetic novelty. Among the five anemones investigated, the distribution of microbial taxa varied; this trait appears to be shared by many anthozoans. Most importantly, specimens either appeared to be dominated by Proteobacteria-affiliated members or by deeply branching Tenericute sequences. There were few closely related sequence types that were common to temperate and Antarctic sea anemone microbiomes, the exception being an Acinetobacter-related representative. Similar observations were made between microbes associated with E. andrillae and an Antarctic soft coral; however, there were several closely-related, low abundance Gammaproteobacteria in both Antarctic microbiomes, particularly from the soft coral, that are also commonly detected in Southern Ocean seawater. Although this preliminary study leaves open many questions concerning microbiome diversity and its role in host ecology, we identify major lineages of microbes (e.g., diverse deep-branching Alphaproteobacteria, Epsilonproteobacteria, and divergent Tenericutes affiliates) that may play critical roles, and we highlight the current understanding and the need for future studies of sea anemone-microbiome relationships.

Analysis of Variance (ANOVA)

Rietveld and van Hout (2005) provide this fictional example involving three second-language vocabulary learning methods (I, II, III), with three different groups of participants assigned to each method.1 The relative effectiveness of the learning methods is evaluated on some scale by scoring the increase in vocabulary after using the method. We draw a sample from each group, and compute the mean scores

The Sampling Distribution of the Sample Mean

We begin by establishing a fundamental fact about any normal distribution: about 95% of the probability lies within 2 SD of the mean. If we integrate the area under these curves, between 2 SD below the mean and 2 SD above the mean, we find the following areas, which correspond to the amount of probability within these bounds

Randomness and Probability

Suppose that, for some reason, we want to know how many times a secondlanguage learner makes errors in a writing task; to be more specific, let’s assume we will only count verb inflection errors. The dependent variable (here, the number of inflection errors) is random in the sense that we don’t know in advance exactly what its value will be each time we assign a writing task to our subject. The starting point for us is the question: What’s the pattern of variability (assuming there is any) in the dependent variable?

Bivariate Statistics and Linear Models

So far we’ve been studying univariate statistics; for example, for each individual in a population, we take a single measurement, height, age, etc. We combine these into a sample and compute a statistic: mean, variance, or some function of the variance. Now we consider the scenario where, for each individual in a population, we have two values: age and height, midterm and final exam result, etc. In such a situation we can, of course, treat each dimension independently, and compute the same univariate statistics as before. But the reason we measure two values is to assess the correlation between them, and for this, we require ‘two-dimensional’ or bivariate statistics.