Skip to contents

AnophelesBionomics

AnophelesBionomics.Rmd

Introduction

This R package provides tools to estimate key bionomics parameters for mosquitoe species using a Bayesian hierarchical model that borrows strength across taxonomic levels (species, species complex, and genus).

We prepare and standardize the input data using taxonomy-aware filtering functions. We then fit a robust Bayesian hierarchical model using rstan. To better understand the observed patterns, we visualize data distributions by species and complex. Finally, we explore the posterior parameter distributions using static density plots generated with ggplot2, as well as interactive charts built with plotly for more intuitive inspection.

Even when species-specific data is missing, the model can infer parameter values using information from related taxa, thanks to taxonomic structure. The approach combines internal datasets included in the package user-provided external data The estimation relies on the No-U-Turn Sampler (NUTS) implemented via rstan, and includes tools for data preparation, posterior diagnostics, and interactive visualization.

The model focuses on estimating the following mosquito-related parameters:

  • parous_rate: Proportion of parous females out of observed females.

  • endophagy: Proportion of mosquitoes biting indoors versus outdoors.

  • endophily: Proportion of mosquitoes resting indoors after having blood-fed.

  • indoor_HBI: Human blood index for mosquitoes collected indoors

  • outdoor_HBI: Human blood index for mosquitoes collected outdoors.

  • sac_rate: Proportion of parous females which have oviposited less than 24 hours before being collected.

  • resting_duration

The necessary and optional packages required to run the package are listed in the appendix.





How to use the package

Adding and cleaning Data

The creation_df() function is designed to preprocess raw entomological data for Bayesian modeling of mosquito bionomic traits. To use it, the user simply specifies the variable of interest — such as Human Blood Index (HBI), endophagy, or endophily — via the varname argument. This variable corresponds to a specific mosquito bionomics

For basic use, only varname is required:

data <- creation_df("endophily")
#> [1] "Total observations: 38"
#> [1] "denominator is always larger or equal to numerator: good!"

When executed, the function returns informative messages in the console : it reports the number of observations retained after filtering (e.g., “Total observations: 90”) and verifies the internal consistency of the data structure by checking that denominators are always greater than or equal to numerators in proportion calculations (e.g., “denominator is always larger or equal to numerator: good!”). These checks help ensure data integrity before downstream analysis.

Optionally, users can apply geographic and temporal filters to refine the dataset. The geo parameter allows restriction to specific regions. Available region codes are “Africa-E” (Eastern Africa), “Africa-W” (Western Africa), “Americas”, “Asia-Pacific”, and NA for unspecified or global studies. The year_min and year_max parameters allow filtering by study period, based on the year of data collection:

data <- creation_df(varname = "parous_rate",
                    geo = c("Africa-W", "Americas"),
                    year_min = 1900,
                    year_max = 2020)
#> [1] "Total observations: 226"
#> [1] "denominator is always larger or equal to numerator: good!"

Internally, the function filters the dataset according to the selected region(s) and time period, extracts the relevant observations for the chosen bionomic parameter, and calculates proportions as needed. It excludes studies not suitable for behavioral inference, handles data from insecticide intervention contexts, and harmonizes mosquito species names. Species are also grouped into complexes when appropriate, enabling hierarchical modeling.

In addition to using internal data, users can also integrate their own external dataset by supplying it through the extern_data argument. This allows appending custom observations to the internal database before preprocessing. The external data must match the internal structure in terms of column names and data types. The rules for formatting this external dataset are provided in the Appendix.

Here’s a minimal working example of external data for the parous_rate trait:

extern_data <- data.frame(
  species = c("Anopheles arabiensis", "Anopheles arabiensis"),
  insecticide_control = c("f", "f"),
  country = c("Benin", "Benin"),
  year_start = c(2010, 2012),
  parity_n = c(45, 30),
  parity_total = c(100, 60),
  parity_percent = c(45, 50),
  stringsAsFactors = FALSE
)

You can include this data in the main function call as follows:

data <- creation_df(varname = "parous_rate",
                    geo = c("Africa-W", "Americas"),
                    year_min = 1900,
                    year_max = 2020, extern_data = extern_data)
#> [1] "Total observations: 228"
#> [1] "denominator is always larger or equal to numerator: good!"

The external data is automatically combined with the internal dataset during preprocessing. If some required columns are missing in extern_data, the function will still run but will exclude those external data rows missing necessary columns, and it will display a message specifying which columns were missing. This allows the internal data to be processed normally while warning the user about incomplete external data.

Make sure your dataset follows the expected structure (see Appendix for a complete column specification) to ensure full integration of your external observations.

The output is a named list containing four elements: (1) data.req, the fully cleaned and filtered dataset, (2) nice_varname, a human-readable label suitable for outputs, (3) species_complex, a reference table mapping species and species complexes to numeric identifiers, and (4) varname, the name of the selected mosquito bionomic parameter.

This function ensures consistent and reproducible preparation of entomological data, focusing on a single mosquito trait per run, and is optimized for integration into Bayesian analysis pipelines.




Data Visualisation

The function obs_complex_species_pie() allows users to visualize the distribution of mosquito observations by species or complex through an interactive pie chart. The function calculates observation proportions and consolidates species with very low representation into an “Other” category based on a user-defined threshold. Additionally, the resulting visualization can be saved as a standalone HTML file if a directory is specified.

When hovering over a slice, detailed information is displayed, including the species or complex name, the total number of observations, the number of survey rows, the mean and standard deviation of the observed values, and the proportion of the total.

This function expects as input a list structured as the output of the creation_df() function, which contains a cleaned data frame with observation counts identified by a variable name pattern, the base name of the variable, and a user-friendly label for output naming. The user can adjust the minimum proportion threshold that determines when rare species are grouped together. For example, setting this threshold to 0.05 means that all categories whose combined total accounts for 5% or less of the observations are grouped into “Other”. Finally the user can chose a path in order to save his pie. For instance, after preparing the data using the preprocessing pipeline, a user might call:

obs_complex_species_pie(data, threshold_prop_other = 0.05)
#> Proportion of the 'Other' category' : 0.03

Similarly, obs_region_pie() generates an interactive pie chart displaying the geographic distribution of mosquito observations. It calculates proportional contributions from regions and labels missing regions as “Unspecified”.

This function also takes as input a list matching the structure returned by creation_df(), ensuring consistent data formatting.For instance, after preparing the data using the preprocessing pipeline, a user might call:

obs_region_pie(data,
                plot_dir = "path/to/save/directory/")

Both functions produce interactive Plotly charts that facilitate exploratory data analysis and can be seamlessly integrated into reporting workflows.




To build the model

The function run_stan() performs hierarchical Bayesian binomial modeling using the rstan package to estimate the distribution of the bionomics, across mosquito species and complexes. It executes the complete modeling workflow, which includes preparing data in a format compatible with Stan via the prepare_stan_data() function, compiling the Stan model, and conducting sampling.

By default, the model runs with 2000 iterations per chain, including a warmup (burn-in) period of 400 iterations, distributed over four independent chains. These parameters — including the total number of iterations, the number of chains, and the thinning interval applied to posterior samples — can be customized by the user to control the length and quality of the sampling process.

The input data must be a cleaned dataset structured as the output from the creation_df() function, containing all necessary variables, including species and complex identifiers, required for model fitting.

Upon completion, the function returns a named list containing several components. The most important of these is the fit element, which is a stanfit object representing the fitted model. This object provides access to posterior distributions of estimated parameters, allowing detailed inspection of species- and complex-level effects. For example, if the function is called as

run_stan_result <- run_stan(data, iter = 1000, chains = 3)
#> 
#> SAMPLING FOR MODEL 'anon_model' NOW (CHAIN 1).
#> Chain 1: 
#> Chain 1: Gradient evaluation took 0.000184 seconds
#> Chain 1: 1000 transitions using 10 leapfrog steps per transition would take 1.84 seconds.
#> Chain 1: Adjust your expectations accordingly!
#> Chain 1: 
#> Chain 1: 
#> Chain 1: Iteration:   1 / 1000 [  0%]  (Warmup)
#> Chain 1: Iteration: 100 / 1000 [ 10%]  (Warmup)
#> Chain 1: Iteration: 200 / 1000 [ 20%]  (Warmup)
#> Chain 1: Iteration: 300 / 1000 [ 30%]  (Warmup)
#> Chain 1: Iteration: 400 / 1000 [ 40%]  (Warmup)
#> Chain 1: Iteration: 500 / 1000 [ 50%]  (Warmup)
#> Chain 1: Iteration: 501 / 1000 [ 50%]  (Sampling)
#> Chain 1: Iteration: 600 / 1000 [ 60%]  (Sampling)
#> Chain 1: Iteration: 700 / 1000 [ 70%]  (Sampling)
#> Chain 1: Iteration: 800 / 1000 [ 80%]  (Sampling)
#> Chain 1: Iteration: 900 / 1000 [ 90%]  (Sampling)
#> Chain 1: Iteration: 1000 / 1000 [100%]  (Sampling)
#> Chain 1: 
#> Chain 1:  Elapsed Time: 3.918 seconds (Warm-up)
#> Chain 1:                0.777 seconds (Sampling)
#> Chain 1:                4.695 seconds (Total)
#> Chain 1: 
#> 
#> SAMPLING FOR MODEL 'anon_model' NOW (CHAIN 2).
#> Chain 2: 
#> Chain 2: Gradient evaluation took 3.7e-05 seconds
#> Chain 2: 1000 transitions using 10 leapfrog steps per transition would take 0.37 seconds.
#> Chain 2: Adjust your expectations accordingly!
#> Chain 2: 
#> Chain 2: 
#> Chain 2: Iteration:   1 / 1000 [  0%]  (Warmup)
#> Chain 2: Iteration: 100 / 1000 [ 10%]  (Warmup)
#> Chain 2: Iteration: 200 / 1000 [ 20%]  (Warmup)
#> Chain 2: Iteration: 300 / 1000 [ 30%]  (Warmup)
#> Chain 2: Iteration: 400 / 1000 [ 40%]  (Warmup)
#> Chain 2: Iteration: 500 / 1000 [ 50%]  (Warmup)
#> Chain 2: Iteration: 501 / 1000 [ 50%]  (Sampling)
#> Chain 2: Iteration: 600 / 1000 [ 60%]  (Sampling)
#> Chain 2: Iteration: 700 / 1000 [ 70%]  (Sampling)
#> Chain 2: Iteration: 800 / 1000 [ 80%]  (Sampling)
#> Chain 2: Iteration: 900 / 1000 [ 90%]  (Sampling)
#> Chain 2: Iteration: 1000 / 1000 [100%]  (Sampling)
#> Chain 2: 
#> Chain 2:  Elapsed Time: 2.748 seconds (Warm-up)
#> Chain 2:                1.523 seconds (Sampling)
#> Chain 2:                4.271 seconds (Total)
#> Chain 2: 
#> 
#> SAMPLING FOR MODEL 'anon_model' NOW (CHAIN 3).
#> Chain 3: 
#> Chain 3: Gradient evaluation took 3.5e-05 seconds
#> Chain 3: 1000 transitions using 10 leapfrog steps per transition would take 0.35 seconds.
#> Chain 3: Adjust your expectations accordingly!
#> Chain 3: 
#> Chain 3: 
#> Chain 3: Iteration:   1 / 1000 [  0%]  (Warmup)
#> Chain 3: Iteration: 100 / 1000 [ 10%]  (Warmup)
#> Chain 3: Iteration: 200 / 1000 [ 20%]  (Warmup)
#> Chain 3: Iteration: 300 / 1000 [ 30%]  (Warmup)
#> Chain 3: Iteration: 400 / 1000 [ 40%]  (Warmup)
#> Chain 3: Iteration: 500 / 1000 [ 50%]  (Warmup)
#> Chain 3: Iteration: 501 / 1000 [ 50%]  (Sampling)
#> Chain 3: Iteration: 600 / 1000 [ 60%]  (Sampling)
#> Chain 3: Iteration: 700 / 1000 [ 70%]  (Sampling)
#> Chain 3: Iteration: 800 / 1000 [ 80%]  (Sampling)
#> Chain 3: Iteration: 900 / 1000 [ 90%]  (Sampling)
#> Chain 3: Iteration: 1000 / 1000 [100%]  (Sampling)
#> Chain 3: 
#> Chain 3:  Elapsed Time: 2.292 seconds (Warm-up)
#> Chain 3:                0.795 seconds (Sampling)
#> Chain 3:                3.087 seconds (Total)
#> Chain 3:
#> Warning: There were 7 divergent transitions after warmup. See
#> https://mc-stan.org/misc/warnings.html#divergent-transitions-after-warmup
#> to find out why this is a problem and how to eliminate them.
#> Warning: Examine the pairs() plot to diagnose sampling problems
#> Warning: Tail Effective Samples Size (ESS) is too low, indicating posterior variances and tail quantiles may be unreliable.
#> Running the chains for more iterations may help. See
#> https://mc-stan.org/misc/warnings.html#tail-ess

Other returned elements include a mapping of species to species complexes, user-friendly and original variable names, and details about the sampling parameters used. These additional outputs are primarily intended to support downstream functions responsible for displaying and further analyzing the model results.




Visualizing Posterior Densities by Species Complex: plot_density()

The plot_density() function is used to visualize the posterior distributions of model parameters after fitting a hierarchical Bayesian model via run_stan(). It generates smooth density plots that summarize the uncertainty associated with each estimated parameter, grouped by complex. These visual summaries provide an intuitive and interpretable view of the model outputs.

The function expects as input the full named list returned by run_stan(), which includes the fitted Stan model object (fit), the mapping between individual species and their associated complexes (species_complex), and a user-friendly variable label (nice_varname) used for titling and file naming. Line styles are used to distinguish levels: dotted lines for genus-level parameters, long-dashed lines for complexes-level, and solid lines for individual species-level. Colors are drawn from a predefined palette, with a fallback to default colors and warnings if certain groups are not defined.

Users can optionally filter the visualization to include only specific species complexes by supplying a vector to the complex_names argument. Parameters labeled with “unlabeled” can be excluded using the “unlabel = TRUE” option. If desired, the final plot can also be saved as a PNG file by specifying a valid path via the plot_dir argument.

For example, assuming the model has already been fitted and stored in an object called stan_result, the density plot can be generated and saved using the following code:

p <- plot_density(stan_results = run_stan_result,
             complex_names = c("Gambiae", "Funestus"),
             unlabel = TRUE)
#> Warning in get_plot_component(plot, "guide-box"): Multiple components found;
#> returning the first one. To return all, use `return_all = TRUE`.
print(p)

The funciton return a combined ggplot object containing posterior density plots by species complex with a shared legend. If plot_dir is specified, the plot is saved as a PNG image file.




Extracting Posterior Summary Statistics

The function species_complex_result() extracts and summarizes posterior statistics from the hierarchical Bayesian model fitted using run_stan(). It produces a tidy data frame ready for analysis or reporting.

The summary includes formatted posterior estimates with confidence intervals, variance measures, and hierarchical metadata. An important feature is the option to set all = TRUE, which enables supplementation of missing species or complexes using the taxonomy. In this case, for species without direct data but associated with a species complex for which data exist, the function uses the complex-level estimates. If no data are available at the complex level, the genus-level estimates are used instead. This hierarchical borrowing of information ensures more complete coverage.

Two flags in the output help track this process: the data flag is set to TRUE if the estimate is derived from original observed data, and FALSE otherwise. The source_flag indicates the level at which the parameter estimate was taken—whether directly from species-level data, species complex, genus, or supplemented externally.

The function expects as input the full named list output from run_stan(), including the fitted Stan model (fit), the species-to-complex mapping (species_complex), and the original variable name (varname). Users can optionally save the results as a CSV file by providing an output directory.

For example, assuming the fitted model results are stored in an object called stan_result, the posterior summaries can be extracted with:

summary_table <- species_complex_result(results = run_stan_result,
                                       all = TRUE)
head(summary_table)
#>                 name   level  estimate    variance  ci_lower  ci_upper data
#> 1          Anopheles   genus 0.5554246 0.024652729 0.2279280 0.8460378 TRUE
#> 2         Albitarsis complex 0.5043656 0.006672325 0.3458019 0.6703306 TRUE
#> 3 Anopheles darlingi complex 0.4996886 0.007277729 0.3409422 0.6754774 TRUE
#> 4 Anopheles moucheti complex 0.6727180 0.005600503 0.5076364 0.7952426 TRUE
#> 5           Funestus complex 0.7161711 0.003448076 0.5719385 0.8070979 TRUE
#> 6            Gambiae complex 0.5924326 0.002889627 0.4695009 0.6911539 TRUE

In order to obtain the correct input for the AnophelsModel package, the function Bionomics_For_Anophles_Model() computes species-specific bionomic parameters using a hierarchical Bayesian framework. It wraps and extends the logic of species_complex_result(), applying it sequentially to the core bionomic traits estimated by the model: parous rate (M, M.sd), endophagy (endophagy, endophagy.sd), endophily (endophily, endophily.sd), sac rate (A0, A0.sd), indoor and outdoor human blood indices, and resting duration.

Human blood feeding behaviour is handled in two steps. Indoor and outdoor human blood indices are first estimated independently by the Bayesian model. The overall human blood index (Chi) is then derived a posteriori as a deterministic function of endophagy. Only the derived HBI (Chi) is retained in the final output.

Internally, for each trait, the function prepares the data using creation_df(), fits the hierarchical Bayesian model with run_stan(), and extracts posterior summaries via the internal helper function species_complex_result_ano_model(). All steps are fully automated: users do not need to run any of these components separately. The resulting output aggregates posterior means and standard deviations across traits into a single tidy data frame.

The function also standardises species and species-complex names to ensure compatibility with the structure expected by AnophelsModel. This includes harmonising known taxonomic variants (e.g. “Anopheles gambiae s.s.” → “Anopheles gambiae” → “Gambiae complex”) and imputing missing species-level values from higher taxonomic levels when required.

The resting duration parameter is estimated by the model and returned under the name tau, preserving consistency with downstream transmission models.

To use the function, simply run:

bionomics_df <- Bionomics_For_Anophles_Model()
head(bionomics_df)

The output format is specifically designed to plug directly into AnophelsModel.




Dashboard

TO DO




Appendix

1. Packages use in AnophelesBionomics

Required Packages

The following packages are required for the core functionalities of the package:

  • stats
  • dplyr
  • ggplot2
  • rstan
  • plotly
  • utils
  • coda

Optional Packages

The following packages are only needed for interactive features:

  • htmlwidgets
  • shiny



2. Species and complexes included

The following is the full list of Anopheles species and species complexes included in the AnophelesBionomics package:

  • Anopheles arabiensis
  • Anopheles funestus complex
  • Anopheles gambiae
  • Anopheles gambiae complex
  • Anopheles funestus
  • Anopheles nili
  • Anopheles moucheti
  • Anopheles nili complex
  • Anopheles melas
  • Anopheles gambiae (Forest)
  • Anopheles gambiae (Form M)
  • Anopheles merus
  • Anopheles nili (ovengensis)
  • Anopheles carnevalei (nili)
  • Anopheles gambiae (Form S)
  • Anopheles gambiae (Savanna)
  • Anopheles gambiae (Bamako)
  • Anopheles gambiae (Mopti)
  • Anopheles albimanus
  • Anopheles darlingi
  • Anopheles pseudopunctipennis complex
  • Anopheles nuneztovari complex
  • Anopheles albitarsis complex
  • Anopheles aquasalis
  • Anopheles albitarsis (marajoara formerly Sp C)
  • Anopheles albitarsis (Sp. E)
  • Anopheles nuneztovari(Sp. B/C)
  • Anopheles albitarsis (formerly Sp A)
  • Anopheles albitarsis (Sp. B)
  • Anopheles quadrimaculatus complex
  • Anopheles quadrimaculatus (formerly Sp. A)
  • Anopheles freeborni
  • Anopheles annularis complex
  • Anopheles culicifacies complex
  • Anopheles aconitus
  • Anopheles subpictus complex
  • Anopheles dirus complex
  • Maculatus Group
  • Anopheles minimus complex
  • Anopheles fluviatilis complex
  • Anopheles barbirostris complex
  • Anopheles maculatus
  • Anopheles barbirostris
  • Anopheles flavirostris
  • Anopheles sundaicus
  • Anopheles sundaicus complex
  • Anopheles sinensis
  • Anopheles harrisoni (formerly minimus Sp. C)
  • Anopheles dirus (formerly Sp. A)
  • Anopheles minimus (formerly Sp. A)
  • Anopheles balabacensis (leucosphyrus)
  • Anopheles stephensi
  • Anopheles anthropophagus (lesteri)
  • Anopheles baimaii (formerly dirus Sp. D)
  • Anopheles leucosphyrus
  • Anopheles cracens (formerly dirus Sp. B)
  • Anopheles fluviatilis (Sp.T)
  • Anopheles culicifacies (Sp. A)
  • Anopheles culicifacies (Sp. B)
  • Anopheles culicifacies (Sp. C)
  • Anopheles fluviatilis (Sp. S)
  • Anopheles culicifacies (Sp. D)
  • Anopheles epiroticus (formerly sundaicus Sp. A)
  • Anopheles annularis
  • Anopheles punctulatus complex
  • Anopheles farauti complex
  • Anopheles koliensis
  • Anopheles latens (leucospyrus)
  • Anopheles leucosphyrus complex
  • Anopheles subpictus (Sp. B)
  • Anopheles farauti (formerly No.1)
  • Anopheles punctulatus
  • Anopheles farauti (No. 4)
  • Anopheles farauti (No. 8)
  • Anopheles fluviatilis (Sp. U)
  • Anopheles annularis (Sp. A)
  • Anopheles culicifacies (Sp. E)
  • Anopheles splendidus
  • Anopheles subpictus
  • Anopheles nivipes
  • Anopheles pseudojamesi
  • Anopheles barbirostris s.l.
  • Anopheles dirus s.l.
  • Anopheles jamesii
  • Anopheles kochi
  • Anopheles maculatus s.l.
  • Anopheles philippinensis
  • Anopheles tessellatus
  • Anopheles vagus
  • Anopheles minimus s.l.
  • Anopheles epiroticus

3. Rules to add data

To successfully create a dataset repository using create_repo(), your dataset must include:

  • species — The mosquito species or complex name.
  • country — The country where the data was collected.
  • year_start — Enables filtering by time periods later.
  • insecticide_control — Should be “f” for control and “t” for intervention.

Optional but recommended:

  • source_id — A unique identifier for the data source or study.
  • citation — Full citation of the study.
  • pubmed_id — PubMed ID related to the study, or NA if not available.

Parous Rate

To calculate parous rate, your dataset must include:

  • parity_n — Number of parous mosquitoes (numerator). Use NA if only percentage is available.
  • parity_total — Total number of mosquitoes dissected (denominator).
  • parity_percent — Percentage; used to estimate parity_n if needed, NA otherwise.
  • insecticide_control — Must be “f”.

endophagy

To calculate endophagy, your dataset must include:

  • indoor_biting_n — Number of mosquitoes biting indoors. Use NA if only percentage is available.
  • indoor_biting_total — Total mosquitoes biting (indoor + outdoor).
  • indoor_biting — Percentage; used to estimate indoor_biting_n if needed, NA otherwise.
  • indoor_biting_sampling — Must be “MBI”.
  • outdoor_biting_sampling — Must be “MBO”.
  • HLC_same — Must be TRUE, indicatting if the estimation used is Human Landing Catches ( adults sit indoors or outdoors with their legs exposed and wait until a mosquito lands on their legs to catch it ).
  • insecticide_control — Must be “f”.

endophily

To calculate endophily, your dataset must include:

  • indoor_fed — Number of mosquitoes resting indoors after having blood-fed.
  • outdoor_fed — Number of mosquitoes resting outdoors after having blood-fed.
  • resting_unit — Must be either “%” or “count”.

If resting_unit is “%”, also include:

  • indoor_total — Total mosquitoes observed indoors.
  • outdoor_total — Total mosquitoes observed outdoors.

Additional required fields:

  • indoor_resting_sampling — Must be “HRI”.
  • outdoor_resting_sampling — Must be “WinExit”.
  • WinExit_in_HRI_houses — Must be TRUE.
  • insecticide_control — Must be “t”.

Optionally, if reconstructing totals:

  • indoor_unfed, indoor_gravid
  • outdoor_unfed, outdoor_gravid

sac rate

To calculate sac rate, your dataset must include:

  • parous_with_sac — of parous females which have oviposited less than 24 hours before being collected.
  • `parous — Total number of parous mosquitoes (denominator).
  • sac_rate_percent — Percentage; used to estimate parous_with_sac if needed, NA otherwise.
  • insecticide_control — Must be “t”.

HBI

To calculate indoor human blood index (indoor HBI), your dataset must include:

  • indoor_host_n — Number of mosquitoes having ingested human blood indoors.
  • indoor_host_total — Number of mosquitoes examined for blood meal origin indoors.
  • indoor_host — Percentage; used to estimate indoor_host_n if needed, NA otherwise.
  • indoor_host_sampling — Must be one of “HRI”, “ILT”, or “WinExit”.
  • host_unit — Must NOT be “AI” (animal index).
  • insecticide_control — Must be “f”.

To calculate outdoor human blood index (outdoor HBI), your dataset must include:

  • outdoor_host_n — Number of mosquitoes having ingested human blood outdoors.
  • outdoor_host_total — Number of mosquitoes examined for blood meal origin outdoors.
  • outdoor_host — Percentage; used to estimate outdoor_host_n if needed, NA otherwise.
  • outdoor_host_sampling — Must be one of “RO”, “RO (shelter)”, “RO (pit)”, or “WinExit”.
  • host_unit — Must NOT be “AI” (animal index).
  • insecticide_control — Must be “f”.

If you have combined host sampling (no separation between indoor and outdoor), your dataset must also, as well that the precedente colomns include:

  • combined_host_n — Number of mosquitoes having ingested human blood.
  • combined_host_total — Number of mosquitoes examined for blood meal origin.
  • combined_host — Percentage; used to estimate combined_host_n if needed, NA otherwise.
  • combined_host_sampling_n — Must NOT be “t”.
  • combined_host_sampling_1 — Main sampling method used, must be valid (different of ““).
  • insecticide_control — Must be “f”.