Aggregation of symbionts on hosts depends on interaction type and host traits
Data files
Sep 13, 2023 version files 73.14 KB

guppygyro.csv

guppysexescorr.csv

limpetbarnacle.csv

README.md
Abstract
Symbionts tend to be aggregated on their hosts, such that few hosts harbor the majority of symbionts. This ubiquitous pattern can result from stochastic processes, but aggregation patterns may also depend on the type of hostsymbiont interaction, plus traits that affect host exposure and susceptibility to symbionts. Untangling how aggregation patterns both within and among populations depend on stochastic processes, interaction type and host traits remains an outstanding challenge. Here, we address this challenge by using null models to compare aggregation patterns in a neutral system of Balanomorpha barnacles attached to patellid limpets and a hostparasite system of Trinidadian guppies (Poecilia reticulata) and their Gyrodactylus spp. monogeneans. We first used a model to predict patterns of symbionthost aggregation due to random partitioning of symbionts to hosts. This null model accurately predicted the aggregation of barnacles on limpets, but the degree of aggregation varied across 303 quadrats. Quadrats with larger limpets had less aggregated barnacles, whereas aggregation increased with variation in limpet size. Across 84 guppy populations, Gyrodactylus spp. parasites were significantly less aggregated than predicted by the null model. As in the neutral limpetbarnacle system, aggregation decreased with mean host size. Parasites were also significantly less aggregated on males than females because male guppies tended to have higher prevalence and lower parasite burdens than predicted by the null model. Together, these results suggest stochastic processes can explain aggregation patterns in neutral but not parasitic systems, though in both systems host traits affect aggregation patterns. Because the distribution of symbionts on hosts can affect symbiont evolution via intraspecific interactions, and reciprocally host behavior and evolution via hostsymbiont interactions, identifying the drivers of aggregation enriches our understanding of hostsymbiont interactions.
README: Aggregation of symbionts on hosts depends on interaction type and host traits
https://doi.org/10.5061/dryad.4b8gthtjx
Here, we test whether the processes that drive aggregation differ between a neutralistic and parasitic hostsymbiont system. We further test how the tolerancelinked sex and exposurelinked size differences of guppy Poecilia reticulata hosts can drive differences in the aggregation of Gyrodactylus spp. parasites among their hosts. For the guppyGyrodactlid portion of the manuscript we are using the guppygyro.csv dataset. This data is a combination of population level data from Stephenson, J. F., Oosterhout, C. van, Mohammed, R. S. & Cable, J. Parasites of Trinidadian guppies: evidence for sex and agespecific traitmediated indirect effects of predators. Ecology 96, 489–498 (2015) and an additional, unpublished data set. For the guppyGyrodactlid portion of the manuscript we are using the guppygyro.csv dataset guppygyro.csv. This data is a combination of population level data from Stephenson, J. F., Oosterhout, C. van, Mohammed, R. S. & Cable, J. Parasites of Trinidadian guppies: evidence for sex and agespecific traitmediated indirect effects of predators. Ecology 96, 489–498 (2015) and an additional, unpublished data set limpetbarnacle.csv. We additionally have a separate CSV file named "guppysexescorr.csv." This dataset is just a subset of the larger guppy dataset to test for correlations between the aggregation of parasites between males and female guppies from the same site.
Description of the data and file structure
guppygyro.csv: contains all data pertaining to the Trinidadian guppy and gyrodactylus parasite data. Included variables:
Site: A factor that identifies the drainage, river, course, site, and year the population data was taken. It has a level per population.
drainage: The drainage where the population was collected from. Six levels
river: The river where the population was collected from. 13 levels.
year: The year the population was collected in. 5 levels.
season: The season the population was collected. 2 levels.
Course: The watercourse the population was collected in. 2 levels. Upper indicates low predation populations while lower indicates high predation populations in line with previous publications from the Trinidadian guppy system
H: The total number of hosts in the population
P: The total number of parasites in the population
MeanP: The mean number of parasites infecting hosts in the population
VarP: The mean number of parasites infecting hosts in the population
MeanLen: The mean length of all individuals in the population (millimeters)
VarLen: The variance in length of the population (millimeters^{2})
MeanW: The mean weight of all individuals in the population (grams)
VarW: The variance in weight of the population. (grams^{2})
fs var log med: The log median variance predicted by the feasible set. \newline
sex: The sex of individuals in the population. Populations were separated into males and females to conduct each analysis and calculate body condition. This allowed us to better understand aggregation differences between sexes. 2 levels.
jonvar: The variation in condition of individuals in the population
ResBC: The body condition given the residual mass index.
jonres: The residual of a regression between the body condition given by index from Johnson et al. 2011 and the mean length of the population. We did this to control for any body condition bias we may see between populations just due to size differences.
diffvar: The difference in variances between the observed and expected distribution (the feasible set) of parasites in the population. Positive values indicates the population has a higher variance (More aggregation) than expected by the feasible set, negative values means the feasible set predicted more variance than is present in the observed data
logMean: The log10 mean of parasites in the population.
jonresRS: The scaled jonres variable from above.
MeanLenRS: The scaled mean length variable from above.
ExpPrev: The expected prevalence of the population given by the feasible set \newline
ObsPrev: The observed prevalence of the population
MaxF: The expected number of parasites on the individual with the highest parasite burden in the population given by the feasible set
MaxE: The observed number of parasites on the individual with the highest parasite burden in the population
We also include an additional guppysexcorr.csv file for an analysis examining the correlation between aggregation metrics between males and females from the same site. A list of the data in this file is included below:
Site: A factor that identifies the drainage, river, course, site, and year the population data was taken. It has a level per population.
FemaleDiff: The difference in variances between the observed and expected distribution (the feasible set) of parasites in the population of females at a given site.
MaleDiff: The difference in variances between the observed and expected distribution (the feasible set) of parasites in the population of males at a given site.
Lastly, we have a file titled limpetbarnacle.csv with all data pertaining to the limpetbarnacle hostsymbiont analysis part of the manuscript. A list of the data in this file is included below:
Site: A factor that identifies the site where the community was collected.
QUADRAT: A factor that identifies the quadrat site location of each community of limpets and barnacles. It has a level per community.
H: The total number of limpet hosts in each community.
P: The total number of barnacle symbionts in each community.
MeanP: The mean abundance of barnacles living on the limpets for each community.
varP: The variance in abundance of barnacles living on the limpets for each population.
MeanW: The mean width measurement for the limpets in each community. Measured in (millimeters)
VarW: The variance in width measurement for the limpets in each community. Measured in (millimeters^{2})
MeanL: The mean length measurement for the limpets in each community. Measured in (millimeters)
VarL: The variance in length measurement for the limpets in each community. Measured in (millimeters^{2})
fs var log med: The log median variance predicted by the feasible set for each limpetbarnacle community.
<br>
<br>