4_transwell_migration_plate_2.Rmd

---
title: "Migration towards Pdgfb Plate 2"
author: "Jordi Camps"
date: "2 Januar 2019"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Load packages
```{r message=FALSE}
library(readxl)
library(tidyr)
library(dplyr)
library(ggplot2)
library(ggpubr)
```

## Load data
```{r}
df <- read_excel("data/Transwell_migration_pdgf_plate_2.xlsx", col_names = TRUE)
df
```

## Create factors
compound
```{r}
df$compound <- plyr::revalue(df$compound, c(assay_buffer = "Negative_control", PDGF_BB = "Pdgfb", thirty_percent_FBS = "Positive_control"))
df$compound <- factor(df$compound, levels = c("Negative_control", "Pdgfb", "Positive_control"))
```

create genotype column
```{r}
df <- df %>%
  separate(celltype, c("celltype", "mouse"), sep = " ")
```

```{r}
df$genotype <- plyr::revalue(df$mouse, c(`WT16/17` = "Healthy", SGCB3 = "Dystrophic"))
df$genotype <- factor(df$genotype, levels = c("Healthy", "Dystrophic"))
head(df)
```

Celltype
```{r}
df$celltype <- factor(df$celltype, levels = c("MAB", "FAP"))
```

## Statistics
```{r}
compare_means(percentage ~ genotype, data = df, group.by = c("celltype", "compound"))
```

## Plot
```{r fig.height=2.5, fig.width=6}
compound_titles <- c(Negative_control = "Negative\ncontrol", Pdgfb = "PDGF-BB", Positive_control = "Positive\ncontrol")
df %>%
  ggplot(aes(x = celltype, y = percentage, fill = genotype)) +
  geom_boxplot() +
  facet_wrap(~compound, scales = "free_y", labeller = labeller(compound = compound_titles)) +
  #scale_y_continuous("Percentage", expand = c(0, 1), limits = c(0, 30)) +
  stat_compare_means(label = "p.signif", label.x = 1) +
  ggthemes::scale_fill_tableau() +
  theme_bw(base_size = 14) +
  theme(panel.border = element_blank(), 
        panel.grid = element_blank(), 
        axis.line = element_line(colour = "black"),
        axis.ticks.y = element_line(colour = "black"), 
        axis.text = element_text(colour = "black"), 
        axis.ticks.x = element_blank(),
        axis.title.x = element_blank(),
        legend.title = element_blank(),
        strip.background = element_blank(),
        strip.text = element_text(face = "italic")
        )
ggsave("plots/migration/percentage_migration_plate_2.pdf", dpi = 600)
ggsave("plots/migration/percentage_migration_plate_2.png", dpi = 600)
```

## Chemotactic index
```{r}
ci <- df %>%
  select(compound, genotype, celltype, percentage) %>%
  group_by(celltype, genotype, compound) %>%
  summarise(replicates = n(),
            perc_avg = mean(percentage, na.rm = TRUE),
            perc_sem = sd(percentage, na.rm = TRUE) / sqrt(replicates) / perc_avg * 100) %>%
  gather(key, value, -(celltype:replicates)) %>%
  unite(col = key, compound, key) %>%
  spread(key, value) %>%
  mutate(CI_pdgfb = Pdgfb_perc_avg / Negative_control_perc_avg,
         SEM_pdgfb = sqrt((Pdgfb_perc_sem)^2 + (Negative_control_perc_sem)^2) * CI_pdgfb / 100,
         CI_posctrl = Positive_control_perc_avg / Negative_control_perc_avg,
         SEM_posctrl = sqrt((Positive_control_perc_sem)^2 + (Negative_control_perc_sem)^2) * CI_posctrl / 100) %>%
  select(celltype, genotype, CI_pdgfb, SEM_pdgfb, CI_posctrl, SEM_posctrl) %>%
  gather(key, value, (CI_pdgfb:SEM_posctrl)) %>%
  separate(key, c("variable", "compound"), sep = "_") %>%
  spread(variable, value) %>%
  mutate(ylower = CI - SEM,
         yupper = CI + SEM)
```

## Plot
```{r fig.height=3, fig.width=4.5}
compound_titles <- c(pdgfb = "Pdgfb", posctrl = "Positive control")
ci %>%
  ggplot(aes(x = celltype, y = CI, fill = genotype)) +
  geom_col(position = position_dodge(), width = 0.8) +
  geom_errorbar(aes(ymin = ylower, ymax = yupper), position = position_dodge(0.8), width = 0.2) +
  geom_hline(yintercept = 1, linetype = 2) +
  #geom_pointrange(aes(ymin = ylower, ymax = yupper), position = position_dodge(0.8)) +
  facet_wrap(~compound, scales = "free_y", labeller = labeller(compound = compound_titles)) +
  scale_y_continuous("Chemotactic index", expand = c(0, 0)) +
  ggthemes::scale_fill_tableau() +
  theme_bw(base_size = 14) +
  theme(panel.border = element_blank(), panel.grid = element_blank(), strip.background = element_blank(), axis.line = element_line(colour = "black"),
        axis.ticks.y = element_line(colour = "black"), axis.text = element_text(colour = "black"), axis.title.x = element_blank(),
        legend.title = element_blank(), strip.text = element_text(face = "italic", colour = "black"), axis.ticks.x = element_blank(),
        legend.position = "bottom")
ggsave("plots/migration/chemotactic_index_plate_2.pdf", dpi = 600)
ggsave("plots/migration/chemotactic_index_plate_2.png", dpi = 600)
```

## Statistics
```{r}
t_data <- list()
for (i in 1:2) {
  for (j in 1:3) {
    temp <- df %>%
      filter(celltype == as.character(unique(df$celltype))[i] & compound == as.character(unique(df$compound))[j])
    t <- t.test(percentage~genotype, data = temp)
    t_data[[3*(i-1)+j]] <- c(p_value = t$p.value, celltype = as.character(unique(df$celltype))[i], compound = as.character(unique(df$compound))[j])
  }
} 
```

```{r}
head(t_data)
```

Create df from statistical list
```{r}
stat <- as.data.frame(t(as.data.frame(t_data, stringsAsFactors = FALSE)), stringsAsFactors = FALSE)
stat$p_value <- as.numeric(stat$p_value)
rownames(stat) <- NULL
head(stat)
```

Merge with ci
```{r}
stat$compound <- plyr::revalue(stat$compound, c(Negative_control = "negctrl", Pdgfb = "pdgfb", Positive_control = "posctrl"))
ci %>%
  inner_join(stat, by = c("celltype", "compound"))
```