In pratikunterwegs/kleptomove-ms: R code for the Kleptomove Simulation

Prepare Activity Budget Plots

Load libraries.

library(data.table)

library(ggplot2)
library(patchwork)
library(colorspace)

devtools::load_all()

Get data

# activity data
data_activity <- fread("data/results/data_pop_activity_budget.csv")
data_activity <- data_activity[
  # sim_type == "foragers" &
    regrowth == 0.01
    # variable != "stealing"
]

# remove stealing from foragers
data_activity = data_activity[!(sim_type == "foragers" & variable == "stealing")]

Split by simulation type.

data_activity = split(
  data_activity, by = "sim_type"
)

Activity budget plot

# fig_activity <-
plot_activity = lapply(data_activity, function(df) {

  # colours and labels
  col_vals = c(
    foraging = "dodgerblue4",
    handling = "forestgreen",
    stealing = "indianred",
    klept = "darkorange"
  )
  label_vals = c(
    foraging = "Searching for prey",
    handling = "Handling prey",
    stealing = "Attempting to steal",
    klept = "% Kleptoparasites"
  )
  if(unique(df$sim_type) %in% c("facultative", "random")) {
    col_vals = col_vals[!names(col_vals) %in% c("klept")]
    label_vals = label_vals[!names(label_vals) %in% c("klept")]
  }
  if(unique(df$sim_type) == c("foragers")) {
    col_vals = col_vals[!names(col_vals) %in% c("stealing", "klept")]
    label_vals = label_vals[!names(label_vals) %in% c("stealing", "klept")]
  }

  ggplot(df) +

  geom_line(
    aes(
      gen, value,
      colour = variable,
      group = interaction(variable, rep)
    )
  ) +
  scale_colour_manual(
    values = col_vals,
    labels = label_vals
    # breaks = c("foraging", "handling", "stealing")
  ) +
  scale_y_continuous(
    labels = scales::percent,
    breaks = seq(0, 1, 0.25)
  ) +
  scale_x_log10(
    breaks = c(1, 3, 10, 30, 100, 300, 1000)
  ) +
  coord_cartesian(
    xlim = c(1, 1000),
    ylim = c(0, 1),
    expand = T
  ) +
  theme_test(
    base_size = 8,
    base_family = "Arial"
  ) +
  theme(
    legend.position = "top",
    legend.key.height = unit(2, units = "mm"),
    axis.text.y = element_text(
      angle = 90, hjust = 0.5
    )
  ) +
  labs(
    x = "Generation",
    y = "% Time",
    colour = NULL
  )+
  guides(colour = guide_legend(nrow = 1))
})

plot_activity = plot_activity[c(
  "foragers", "obligate", "facultative", "random"
)]

Proportion of fixed kleptoparasites

# load the relative preference evolution dataset
data_strat = fread("data/results/data_relative_preference_evolution.csv")

# keep only the data for wt_5, aka w0, the bias to be a kleptoparasite
# examine only scenario 2
data_strat = data_strat[wt == "wt_5" & sim_type == "obligate",]

# assign strategy and sum proportions
data_strat[, strategy := fifelse(wt_value > 0, "forager", "klept")]

data_strat = data_strat[, list(
  prop = sum(prop)
), by = c("gen", "rep", "strategy")]

# add the proportion of the population that are fixed kleptoparasites
# as an element of the plot for scenario 2
plot_activity[["obligate"]] = plot_activity[["obligate"]] +
  geom_path(
    data = data_strat[strategy == "klept"],
    aes(
      gen, prop, col = strategy, 
        group = interaction(strategy, rep)
    ),
    position = "identity"
  )

Mean per capita intake plot

plot_intake <- lapply(data_activity, function(df) {
  ggplot(
    unique(df,
      by = c("gen", "rep", "pop_fitness")
    )
  ) +
  geom_path(
    aes(gen, pop_fitness,
      group = rep
    ),
  ) +
  scale_x_log10(
    breaks = c(1, 3, 10, 30, 100, 300, 1000)
  ) +
  coord_cartesian(
    # xlim = c(1, 50),
    xlim = c(1, 1000),
    ylim = c(0, 50),
    expand = T
  ) +
  theme_test(
    base_size = 8, 
    base_family = "Arial"
  ) +
  labs(
    x = "Generation",
    y = "Mean per capita intake"
  )
})

plot_intake = plot_intake[c(
  "foragers", "obligate", "facultative", "random"
)]

Correlation with quality

# raw correlation data
data_quality <- fread("data/results/data_corr_input_matching.csv")
data_quality <- data_quality[regrowth == 0.01, ]

data_quality = split(data_quality, by = "sim_type")

plots_quality = lapply(data_quality, function(df) {

  ggplot(df) +
  geom_hline(
    yintercept = 0,
    lty = 2,
    size = 0.2
  ) +
  geom_point(
    data = df,
    aes(
      x = gen,
      y = cf
    ),
    shape = 1,
    colour = "navy",
    stroke = 0.5,
    show.legend = length(unique(df$strategy)) > 1
  ) +
  scale_x_log10(
    breaks = c(1, 3, 10, 30, 100, 300, 1000)
  ) +
  theme_test(
    base_size = 8,
    base_family = "Arial"
  ) +
  theme(
    legend.key.height = unit(2, units = "mm")
  )+
  coord_cartesian(
    ylim = c(-0.5, 0.5),
    xlim = c(1, 1000),
    expand = T
  ) +
  labs(
    x = "Generation",
    y = "Corr. # indivs. ~ cell quality",
    colour = NULL,
    shape = NULL
  )
})

plots_quality = plots_quality[c(
  "foragers", "obligate", "facultative", "random"
)]

Plot landscape and agent distributions

# get landscape and agent data
data_land <- fread("data/results/data_landscape_progression.csv")
data_land <- data_land[gen %in% c(1, 10, 50, 950) & regrowth == 0.01, ]

data_agent = fread("data/results/data_agent_locations.csv")
data_agent <- data_agent[gen %in% c(1, 10, 50, 950) & regrowth == 0.01, ]

data_land = split(data_land, by = "sim_type")
data_agent = split(data_agent, by = "sim_type")

plots_landscape = Map(
  data_land, data_agent, 

  f = function(df_land, df_agent) {
  ggplot() +
  geom_tile(
    data = df_land,
    aes(x, y, fill = items),
    alpha = 0.5
  ) +
  geom_point(
    data = df_agent[agents > 0, ],
    aes(x, y, colour = agents),
    shape = 4,
    size = 0.5,
    stroke = 0.5
  ) +
  facet_grid(
    ~gen,
    labeller = label_both
  ) +
  scale_fill_continuous_sequential(
    palette = "Batlow",
    l1 = 20, l2 = 100,
    c1 = 10,
    begin = 0.2,
    limits = c(1, NA),
    na.value = "white",
    name = "# Prey",
    rev = T,
    guide = guide_legend(order = 1)
  )+
  scale_colour_continuous_sequential(
    palette = "Reds 2",
    begin = 0.5,
    na.value = "black",
    limits = c(1, 5),
    name = "# Consumers",
    guide = guide_legend(order = 2)
  )+
  coord_equal(expand = F) +
  theme_custom(
    landscape = T, 
    base_size = 8,
    base_family = "Arial"
  ) +
  theme(legend.position = "bottom")
})

plots_landscape = plots_landscape[c(
  "foragers", "obligate", "facultative", "random"
)]

Relative strategy preferences

# load relative preferences
data_pref = fread("data/results/data_rel_pref.csv")
data_pref = data_pref[rep == 3,]

# split by sim type
data_pref = split(data_pref, by = "sim_type")

plot_pref = Map(
  data_pref, names(data_pref),

  f = function(df, n) {

    # generations to plot
    g = c(10, 30, 300, 950)

    # which generations if scenario 3
    if(n == "facultative") g = c(10, 30, 950)

    p = ggplot(df[gen %in% g])+
      geom_abline(
        intercept = c(0, 1, -1),
        slope = c(0, -1, 1),
        lty = 2,
        size = 0.2,
        col = "grey"
      )+
      geom_jitter(
        aes(sP, sH, fill = sN),
        colour = "grey50",
        shape = 21,
        # stroke = 0.2,
        alpha = 0.5,
        size = 2
      )+
      scale_fill_continuous_diverging(
        palette = "Blue-Red 2",
        rev = T,
        limits = c(-1, 1),
        breaks = c(-1, 1),
        labels = c("Avoid", "Prefer")
      )+
      scale_x_continuous(
        breaks = c(0, 1),
        labels = c("Neutral", "Prefer")
      )+
      scale_y_continuous(
        breaks = c(-1, 1),
        labels = c("Avoid", "Prefer")
      )+
      coord_cartesian(
        xlim = c(0, 1),
        ylim = c(-1, 1)
      )+
      theme_test(
        base_size = 8,
        base_family = "Arial"
      )+
      theme(
        axis.text.y = 
        element_text(
          hjust = c(0, 1),
          angle = 90
        ),
        axis.text.x = 
        element_text(
          hjust = c(0, 1)
        ),
        strip.text = element_text(
          face = "italic"
        ),
        strip.background = element_blank(),
        legend.position = "top",
        legend.title = element_text(vjust = 1.5),
        legend.key.height = unit(1, "mm"),
        legend.key.width = unit(5, "mm")
      )+
      labs(
        x = "sP: Prey-item preference",
        y = "sH: Handler preference",
        fill = "sN: Non-handler preference"
      )

      if (n == "obligate") {
        p = p +
          facet_grid(
            rows = vars(comp_strat),
            cols = vars(gen),
            labeller = labeller(
              comp_strat = c(
                "forager" = "Foragers",
                "klept" = "Klept."
              ),
              gen = function(x) sprintf("Gen = %s", x)
            )
          )
      } else {
        p = p +
          facet_grid(
            cols = vars(gen),
            labeller = labeller(
              gen = function(x) sprintf("Gen = %s", x)
            )
          )
      }
      p
  }
)

plot_pref = plot_pref[c(
  "foragers", "obligate", "facultative", "random"
)]

Conditional strategies: Competition choices under ecological scenarios

data_comp = fread("data/results/data_sc3_comp_choice.csv")

data_comp = data_comp[gen == 950,]
data_comp = data_comp[, list(
  prop_steal = sum(steal) / (2500*3)
), by = c("food", "handlers", "gen")]

fig_comp = ggplot(data_comp)+
  geom_tile(
    aes(
      food, handlers, 
      fill = prop_steal
    )
  )+
  scale_fill_continuous_sequential(
    palette = "Purple-Yellow",
    # limits = c(0, 1),
    breaks = c(0.4, 0.7, 1),
    labels = scales::percent_format(accuracy = 1)
    # trans = "sqrt",
  )+
  scale_x_continuous(
    breaks = seq(0, 5)
  )+
  scale_y_continuous(
    breaks = seq(0, 5)
  )+
  facet_grid(
    cols = vars(gen),
    labeller = labeller(
      gen = function(x) sprintf("Gen = %s", x)
    )
  )+
  coord_cartesian(
    expand = F
  )+
  theme_test(
    base_size = 8,
    base_family = "Arial"
  )+
  theme(
    legend.position = "top",
    legend.key.height = unit(1, "mm"),
    legend.key.width = unit(4, "mm"),
    legend.title = element_text(
      vjust = 1.2
    ),
    strip.text = element_text(
      face = "italic"
    ),
    strip.background = element_blank(),
  )+
  labs(
    x = "# Prey-items",
    y = "# Handlers",
    fill = "% Consumers\nstealing"
  )
fig_comp

Make scenario specific figures

figs_1_2_4 = Map(
  plots_landscape, plot_activity, plot_intake, plots_quality,
  plot_pref, names(plot_pref),
  f = function(p1, p2, p3, p4, p5, n) {

    # handle common elements of all plots
    p = wrap_plots(
      p2, p3, p4,
      design = "ABC"
    ) +
    plot_layout(guides = "collect") &
    theme(
      legend.position = "bottom"
    )

    # handle scenario 2 plot
    if(n %in% c("foragers", "random")) {
      p = wrap_plots(
        p1,
        p,
        p5,
        ncol = 1
      ) +
      plot_annotation(
        tag_levels = "A"
      ) &
      theme(
        plot.tag = element_text(
          face = "bold",
          size = 10
        )
      )
    } else if (n == "facultative") {
       p = wrap_plots(
         p1,
         p,
         p5,
         fig_comp,
         design = "AAA\nBBB\nCCD"
       )+
      plot_annotation(
        tag_levels = "A"
      ) &
      theme(
        plot.tag = element_text(
          face = "bold",
          size = 10
        )
      )
    } else {
      p = wrap_plots(
        p1,
        p,
        ncol = 1
      ) +
      plot_annotation(
        tag_levels = "A"
      ) &
      theme(
        plot.tag = element_text(
          face = "bold",
          size = 10
        )
      )
    }
    p
  }
)

# check names
names(figs_1_2_4)

# save figures
invisible(
  Map(figs_1_2_4, c(as.character(c(1,2,4)), "random"), f = function(p, n) {
    if(n == "2") {
      ggsave(
        p,
        filename = sprintf("supplement/figures/fig_0%s.png", n),
        height = 110, width = 150, units = "mm"
      )
    } else {
      ggsave(
        p,
        filename = sprintf("figures/fig_0%s.png", n),
        height = 150, width = 150, units = "mm"
      )
    }

  })
)