NetworksSummer2021-ColonyMovement.Rmd

---
title: "Colony Movement Analysis"
author: "Matina Donaldson-Matasci"
date: "2024-05-07"
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
require(tidyverse)
require(broom)
require(ggthemes)
require(grDevices)
require(lme4)
require(glmmTMB)
require(bbmle) #for AICtab

# theme_set(theme_test(base_size=18)) # For poster
# theme_set(theme_test(base_size=10)) # For paper
theme_set(theme_test(base_size=14)) # For talk


nest.palette <- palette.colors(n = 9, palette = "Classic Tableau")[c(1:7,9)]
nest.names <- c("A2","B3","C1","D2","E1","F1","G1","H1")
names(nest.palette) = nest.names
```

## Hypotheses / Questions

We would like to explore the movement and turning choices of individual ants moving through a branch maze when they are in a colony context. Some questions:

1. Do the turning choices (left, right, u-turn) at individual junctions match those that were observed in the individual movement trials? If not, are they more extreme? This could indicate positive feedback.
2. How do the turning choices at individual junctions change over time? Do they become more extreme? This is another, stronger indicator of positive feedback.
3. What is the proportion of arrivals at each nest? Do the overall numbers match what's expected from the individual movement trials? If not, are they more concentrated at specific nests?
4. What is the temporal pattern of arrivals at each nest? Are arrivals temporally autocorrelated?

## Nest arrivals

First, let's load in the colony-level tracking data. This gives information on all the turns taken within the first 30 minutes of the colony experiment.

```{r}
BA2 <- read.csv("data/BA2-Colony-90min-edges.csv") 
SP2 <- read.csv("data/SP2-Colony-90min-edges.csv")
MW3 <- read.csv("data/MW3-Colony-90min-edges.csv") 
SP3 <- read.csv("data/SP3-Colony-90min-edges.csv")

colony_turns <- bind_rows("21-BA2" = BA2, "21-SP2" = SP2, 
                          "21-MW3" = MW3, "21-SP3" = SP3, 
                          .id="Colony") %>%
  mutate(Colony = factor(Colony, 
                            levels= c("21-MW3","21-BA2","21-SP3","21-SP2"),
                            ))

colony_turns <- colony_turns %>%
  mutate(Junction = as.numeric(str_sub(roi, 5, -1))) %>%
  mutate(Direction.From = case_when(
    edge0 == "Left" ~ "left",
    edge0 == "Right" ~ "right",
    edge0 == "Base" ~ "main"
  )) %>%
  mutate(Direction.To = case_when(
    edge1 == "Left" ~ "left",
    edge1 == "Right" ~ "right",
    edge1 == "Base" ~ "main"    
  ))

turns <- read.csv("data/NetworksSummer2021-Turns.csv")
colony_turn_choices <- colony_turns %>%
  left_join(turns, 
    by=c("Junction", "Direction.From", "Direction.To"))
```

Now let's just filter out turns where ants are headed toward a nest. We can use this to plot a running tally of arrivals at each nest over time.

```{r}
tip_ref <- read.csv("data/NetworksSummer2021-TipReference.csv")

colony_nest_arrivals <- colony_turn_choices %>%
  left_join(tip_ref,
            by=c("Node.To" = "Tip")) %>%
  filter(Type=="nest") %>%
  group_by(Colony,Location) %>%
  mutate(Tally=1) %>%
  mutate(Cumulative.Arrivals = cumsum(Type=="nest"))

colony_nest_arrivals %>% 
  ggplot(aes(x=t1,y=Cumulative.Arrivals,color=Location)) +
  geom_line() +
  scale_color_manual(values=nest.palette,drop=F) +
  xlab("Time after experiment start (min)") +
  ylab("Cumulative arrivals at nest") +
  facet_wrap(. ~ Colony, scales="free_y")
```

We could also try to include the nest departures to get a kind of running tally.

```{r}
colony_nest_departures <- colony_turn_choices %>%
  left_join(tip_ref,
            by=c("Node.From" = "Tip")) %>%
  filter(Type=="nest") %>%
  group_by(Colony,Location) %>%
  mutate(Tally=-1) %>%
  mutate(Cumulative.Departures = -cumsum(Type=="nest"))

colony_nest_tally <- bind_rows(colony_nest_arrivals, 
                               colony_nest_departures) %>%
  group_by(Colony) %>%
  arrange(t0, .by_group=TRUE) %>%
  group_by(Colony,Location) %>%
  mutate(Cumulative.Tally = cumsum(Tally))

colony_nest_tally %>% 
  ggplot(aes(x=t1,y=Cumulative.Tally,color=Location)) +
  geom_line() +
  scale_color_manual(values=nest.palette,drop=F) +
  xlab("Time after experiment start (min)") +
  ylab("Cumulative tally at nest") +
  facet_wrap(. ~ Colony, scales="free_y")

```
This doesn't look great, we should not have negative numbers here. Hmmm.

```{r}
colony_tip_arrivals <- colony_turn_choices %>%
  inner_join(tip_ref,
            by=c("Node.To" = "Tip")) %>%
  group_by(Colony,Location) %>%
  mutate(Type = case_when(Type=="empty" ~ "other",
                          Type=="tag" ~ "other",
                          .default = Type)) %>%
  mutate(Type = factor(Type, 
                       levels = c("nest", "trunk", 
                                  "water", "sugar", 
                                  "other"))) %>%
  separate_wider_position(Location, widths=c(Section=1),
                          too_many="drop", 
                          cols_remove=FALSE) %>%
  mutate(Tally=1) %>%
  mutate(Cumulative.Arrivals = cumsum(Tally))

#tip.palette <- rep("grey", nrow(tip_ref))
#tip.palette <- rep(nest.palette, each=4)
section.palette <- nest.palette
names(section.palette) <- LETTERS[0:8]
#tip.palette[names(nest.palette)] <- nest.palette
tip.linetypes <- c(nest=1,water=2,sugar=3,other=4,trunk=6)
tip.linewidths <- c(nest=1,water=0.5,sugar=0.5,other=0.5,trunk=1)


colony_tip_arrivals %>% 
  ggplot(aes(x=t1,y=Cumulative.Arrivals,color=Section,
             linetype = Type, linewidth = Type)) +
  geom_line() +
  scale_color_manual(values=section.palette) +
  scale_linetype_manual(values=tip.linetypes) +
  scale_linewidth_manual(values=tip.linewidths) +
  xlab("Time after experiment start (min)") +
  ylab("Cumulative arrivals at branch tips") +
  facet_grid(Colony~., scales="free_y")

ggsave("images/ColonyChoice-Tracking90min-TALK.eps", 
       width = 6.5, height = 7, units = "in")
```

Let's try to plot what's happening at all the different junctions.

```{r}
colony_turn_choices <- colony_turn_choices %>% 
  mutate(
    Branch.From = case_when(
      Direction.From == "main" ~ "main",
      Direction.From == "left" & Junction.Handedness == "LH" ~ "secondary",
      Direction.From == "left" & Junction.Handedness == "RH" ~ "primary",
      Direction.From == "right" & Junction.Handedness == "LH" ~ "primary",
      Direction.From == "right" & Junction.Handedness == "RH" ~ "secondary"
    )
  ) %>%
  group_by(Colony, roi, Branch.From) %>%
  mutate(
    Angle.Numeric = case_when(
      Turn.Angle == "shallow" ~ 1, 
      Turn.Angle == "sharp" ~ -1, 
      .default = 0), 
    Angle.Tally = cumsum(Angle.Numeric)
    ) %>%
  separate_wider_delim(roi, "_", names=c(NA,"roi.Number"),
                       cols_remove=FALSE) %>%
  mutate(roi.Section = case_when(
    str_starts(roi.Number, "3") ~ "A",
    str_starts(roi.Number, "5") ~ "B",
    str_starts(roi.Number, "7") ~ "B",
    str_starts(roi.Number, "11") ~ "C",
    str_starts(roi.Number, "12") ~ "D",
    str_starts(roi.Number, "4") ~ "E",
    str_starts(roi.Number, "8") ~ "F",
    str_starts(roi.Number, "6") ~ "F",
    str_starts(roi.Number, "21") ~ "G",
    str_starts(roi.Number, "22") ~ "H",
    .default = "none"
  )) %>%
  mutate(roi.Depth = as.character(str_length(roi.Number)))

section.palette = c("Grey",nest.palette)
names(section.palette) = c("none", LETTERS[0:8])

colony_turn_choices %>%
  ggplot(aes(x=t0, y=Angle.Tally, color=roi.Section, linetype=roi.Depth)) +
  geom_line(aes(group=roi.Number)) +
  scale_color_manual(values=section.palette, drop=F) +
  facet_grid(Colony ~ Branch.From, scales="free_y")
```
```{r}
colony_turn_choices %>%
  filter(Colony == "21-SP2") %>%
  ggplot(aes(x=t0, y=Angle.Tally, color=Branch.From)) +
  geom_line() +
  facet_wrap(vars(roi.Number))
```