parasiteR

A Theoretical-Practical Approach to Parasitological Data Analysis

parasiteR is an R package designed to standardize and streamline the analysis of parasitological data. It provides tools for importing, exploring, summarizing, and visualizing parasite abundance data, as well as estimating key parasitological descriptors and their confidence intervals.

The package is particularly useful for researchers working with parasite–host datasets and quantitative parasitology, providing a workflow from raw CSV data to statistical summaries and graphical representations.

Main features

parasiteR provides functions for:

• Importing and checking parasitological datasets from .csv files.
• Exploring parasite abundance and prevalence patterns.
• Calculating common parasitological descriptors.
• Estimating parasite abundance and confidence intervals.
• Estimating parasite prevalence and confidence intervals.
• Estimating parasite intensity using mean or median measures.
• Quantifying parasite aggregation using several aggregation indices.
• Calculating confidence intervals using bootstrap and likelihood-profile approaches.
• Visualizing parasitological estimates and their confidence intervals.
• Exporting statistical results and figures directly from the analysis workflow.

Installation

Development version from GitHub

You can install the development version of parasiteR from GitHub using remotes:

# Install remotes if necessary
install.packages("remotes")

# Install parasiteR
# remotes::install_github("Dvelop-R/parasiteR")
install.packages("parasiteR", repos = c('https://dvelop-r.r-universe.dev', 'https://cloud.r-project.org'))

Load the package

library(parasiteR)

Data format

The package is designed to work with datasets in which each row represents an individual host.

Columns can include:

Host-related variables

These variables describe the host or sampling context, for example:

• Sampling site
• Host species
• Sex
• Morphophysiological characteristics
• Experimental treatment
• Other categorical or numerical metadata

Parasite-related variables

Parasite abundance data should be represented by one or more columns, with each column corresponding to a parasite taxon.

For example:

Site	Sex	Sp1	Sp2	Sp3
A	F	0	4	2
A	M	3	0	1
B	F	8	2	0
B	M	NA	5	3

Parasite abundance values should be non-negative integers.

It is important to distinguish between:

• 0: The parasite was not detected in the host.
• NA: Parasite detection or quantification was not possible or the observation is missing.

This distinction is important for correctly estimating prevalence, abundance, and other parasitological descriptors.

Included example dataset

The package includes an example dataset called para_data.

You can inspect it with:

data("para_data")

str(para_data)

The dataset can be used to test the main functions and reproduce examples from the package documentation.

Workflow

A typical parasiteR workflow can be summarized as:

Raw parasitological data
        |
        v
   para_read_data()
        |
        v
Data exploration
        |
        +--------------------+
        |                    |
        v                    v
Abundance / Prevalence   Descriptive indices
        |                    |
        v                    v
Confidence intervals    Aggregation measures
        |                    |
        +---------+----------+
                  |
                  v
          para_plot_CI()
                  |
                  v
       Tables and visualizations

Reading parasitological data

The para_read_data() function imports CSV files and provides information about the resulting dataset.

data <- para_read_data(
  file_name = "my_parasite_data.csv",
  verbose = TRUE
)

data$dataset
data$factors_v
data$num_v
data$summ

The function returns a list containing:

• dataset: The imported data frame.
• factors_v: Names of categorical/factor variables.
• num_v: Names of numerical variables.
• summ: A summary of the dataset.

Both comma-separated and semicolon-separated CSV files are supported.

Exploring parasite abundance

The package includes exploratory functions for examining parasite abundance and prevalence patterns.

Abundance exploration

para_explo_abund(
  dataset = para_data$dataset,
  sp_cols = c("Sp1", "Sp2")
)

Prevalence exploration

para_explo_prev(
  dataset = para_data$dataset,
  sp_cols = c("Sp1", "Sp2")
)

These functions can be used as an initial step to explore the structure of parasitological data before calculating summary statistics and confidence intervals.

Parasitological descriptors

The para_descriptors() function calculates descriptive parasitological measures from parasite abundance data.

descriptors <- para_descriptors(
  dataset = para_data$dataset,
  sp_cols = c("Sp1", "Sp2"),
  group_vars = c("Site")
)

descriptors

Depending on the analysis, the package supports descriptors related to parasite abundance, prevalence, intensity, and aggregation.

Abundance and confidence intervals

The para_abundance_CI() function estimates parasite abundance and associated confidence intervals.

abundance_CI <- para_abundance_CI(
  dataset = para_data$dataset,
  sp_cols = c("Sp1", "Sp2"),
  group_vars = c("Site"),
  c_median = FALSE,
  perm = 2000,
  ci_method = "bca",
  conf_level = 0.95,
  combine_ci = FALSE,
  verbose = TRUE
)

abundance_CI

The function can be used to estimate mean or median abundance, depending on the selected options.

Confidence intervals can be estimated using bootstrap procedures, including:

• BCa (bias-corrected and accelerated) bootstrap intervals.
• Percentile bootstrap intervals.

Parasite prevalence

para_prevalence_CI() estimates parasite prevalence and its confidence interval.

prevalence_CI <- para_prevalence_CI(
  dataset = para_data$dataset,
  sp_cols = c("Sp1"),
  group_vars = c("Site"),
  decimal_places = 2,
  conf_level = 0.95,
  ci_method = "wilson",
  output_type = "proportion",
  combine_ci = TRUE,
  verbose = TRUE
)

prevalence_CI

Available confidence interval methods include:

• "wilson" — Wilson score interval.
• "exact" — Clopper–Pearson exact binomial interval.
• "blaker" — Blaker exact interval.

Prevalence can be returned as either a proportion or percentage:

output_type = "proportion"

or

output_type = "percentage"

Parasite intensity

The para_intensity_CI() function estimates parasite intensity among infected hosts.

Both mean and median intensity can be calculated.

Median intensity

median_intensity <- para_intensity_CI(
  dataset = para_data$dataset,
  c_median = TRUE,
  sp_cols = c("Sp1"),
  group_vars = c("Site"),
  perm = 2000,
  ci_method = "bca",
  conf_level = 0.95,
  combine_ci = TRUE,
  verbose = TRUE
)

Mean intensity

mean_intensity <- para_intensity_CI(
  dataset = para_data$dataset,
  c_median = FALSE,
  sp_cols = c("Sp1"),
  group_vars = c("Site"),
  perm = 2000,
  ci_method = "bca",
  conf_level = 0.95,
  combine_ci = TRUE,
  verbose = TRUE
)

Bootstrap methods are used to estimate confidence intervals without requiring a specific parametric distribution.

This approach is particularly useful for parasitological data, which are often highly skewed, overdispersed, and aggregated among hosts.

Parasite aggregation

para_aggregation_CI() calculates several measures used to describe parasite aggregation.

aggregation_CI <- para_aggregation_CI(
  dataset = para_data$dataset,
  sp_cols = c("Sp1", "Sp2"),
  group_vars = c("Site"),
  perm = 2000,
  decimal_places = 2,
  ci_method = "bca",
  combine_ci = FALSE,
  conf_level = 0.95,
  verbose = TRUE
)

aggregation_CI

The function provides estimates for several aggregation measures, including:

• Variance-to-Mean Ratio (VMR)
• Mean crowding
• Patchiness index
• Poulin's D
• Hoover's index
• Negative binomial k

Confidence intervals for most aggregation measures are estimated using bootstrap methods. The negative binomial k parameter uses likelihood-profile confidence intervals.

Available bootstrap confidence interval methods include:

• "bca" — Bias-corrected and accelerated bootstrap.
• "perc" — Percentile bootstrap.

Visualization

The para_plot_CI() function creates graphical representations of parasitological estimates and their confidence intervals.

It is designed to work directly with the outputs of:

• para_abundance_CI()
• para_intensity_CI()
• para_prevalence_CI()
• para_aggregation_CI()

For example:

para_plot_CI(
  para_df = prevalence_CI,
  group_vars = "Site",
  sp_cols = "Sp1"
)

Multiple grouping variables can also be used:

para_plot_CI(
  para_df = prevalence_CI,
  group_vars = c("Site", "Sex"),
  sp_cols = c("Sp1", "Sp2")
)

The function automatically detects the appropriate estimate and confidence interval structure from the original output.

Plots can be displayed as:

• A single faceted plot.
• Separate plots for each parasite taxon.

Figures can also be exported in:

• PDF
• JPEG
• TIFF
• SVG

For example:

para_plot_CI(
  para_df = prevalence_CI,
  group_vars = "Site",
  save_fig = "pdf"
)

Complete example

A basic analysis workflow can be performed as follows:

library(parasiteR)

# Load example data
data("para_data")

# Select parasite taxa
species <- c("Sp1", "Sp2")

# Calculate prevalence
prev <- para_prevalence_CI(
  dataset = para_data$dataset,
  sp_cols = species,
  group_vars = "Site",
  ci_method = "wilson",
  conf_level = 0.95,
  combine_ci = TRUE
)

# Inspect results
print(prev)

# Visualize prevalence and confidence intervals
para_plot_CI(
  para_df = prev,
  group_vars = "Site",
  sp_cols = species
)

Statistical methods

The package implements methods commonly used in quantitative parasitology.

The main methodological approaches include:

• Descriptive statistics for parasite abundance.
• Binomial confidence intervals for prevalence.
• Wilson score intervals.
• Clopper–Pearson exact intervals.
• Blaker exact intervals.
• Non-parametric bootstrap confidence intervals.
• Bias-corrected and accelerated (BCa) bootstrap intervals.
• Percentile bootstrap intervals.
• Likelihood-profile confidence intervals for the negative binomial k parameter.

Because parasitological data frequently exhibit zero inflation, overdispersion, skewness, and aggregation, non-parametric approaches are used where appropriate.

Users should consider sample size and the biological context of their data when interpreting estimates and confidence intervals. Very small sample sizes may result in unstable or uninformative estimates.

Functions

Function	Purpose
para_read_data()	Import and summarize parasitological CSV data
para_explo_abund()	Explore parasite abundance
para_explo_prev()	Explore parasite prevalence
para_descriptors()	Calculate parasitological descriptors
para_abundance_CI()	Estimate parasite abundance and confidence intervals
para_prevalence_CI()	Estimate prevalence and confidence intervals
para_intensity_CI()	Estimate mean or median parasite intensity and confidence intervals
para_aggregation_CI()	Estimate parasite aggregation indices and confidence intervals
para_plot_CI()	Visualize estimates and confidence intervals

Dependencies

parasiteR relies on several R packages, including:

• dplyr
• tidyr
• ggplot2
• rlang
• magrittr
• readr
• stats
• boot
• binom
• BlakerCI
• bbmle
• MASS

Authors

Juan Manuel Cabrera Creator, maintainer, and author

Exequiel Oscar Furlan Author

Elisa Helman Author

References

Bush, A. O., Lafferty, K. D., Lotz, J. M., & Shostak, A. W. (1997). Parasitology meets ecology on its own terms: Margolis revisited. Journal of Parasitology, 83(4), 575–583. https://doi.org/10.2307/3284227

Reiczigel, J., Marozzi, M., Fábián, I., & Rózsa, L. (2019). Biostatistics for parasitologists – a primer to quantitative parasitology. Trends in Parasitology, 35(4), 277–281. https://doi.org/10.1016/j.pt.2019.01.003

Morrill, A., Poulin, R., & Forbes, M. (2023). Interrelationships and properties of parasite aggregation measures: A user's guide. International Journal for Parasitology, 53, 763–776.

License

parasiteR is released under the GNU General Public License version 3 (GPL-3).

Citation

If you use parasiteR in your research, please cite the package and the relevant methodological references used for your analysis.

---