How to combine multiple plots in R with cowplot and ggplot2 (CC098)

#schubert_diversity.R

library(tidyverse)
library(readxl)
library(glue)
library(ggtext)

set.seed(19760620)

metadata <- read_excel(path="raw_data/schubert.metadata.xlsx", na="NA") %>%
  mutate(disease_stat = factor(disease_stat,
                               levels=c("NonDiarrhealControl",
                                        "DiarrhealControl",
                                        "Case")
                               )
  )

alpha_diversity <- read_tsv("raw_data/schubert.groups.ave-std.summary") %>%
  filter(method == "ave") %>%
  select(-label, -method)

metadata_alpha <- inner_join(metadata, alpha_diversity,
                             by=c('sample_id'='group')
                             )

healthy_color <- "#BEBEBE"
diarrhea_color <- "#0000FF"
case_color <- "#FF0000"

disease_count <- metadata_alpha %>%
  count(disease_stat)

healthy_n <- disease_count %>%
  filter(disease_stat == "NonDiarrhealControl") %>%
  pull(n)

diarrhea_n <- disease_count %>%
  filter(disease_stat == "DiarrhealControl") %>%
  pull(n)

case_n <- disease_count %>%
  filter(disease_stat == "Case") %>%
  pull(n)


kt <- kruskal.test(invsimpson ~ disease_stat, data=metadata_alpha)

if(kt$p.value < 0.05){
  pt <- pairwise.wilcox.test(metadata_alpha$invsimpson,
                       g=metadata_alpha$disease_stat,
                       p.adjust.method = "BH")
}

strip_chart <- metadata_alpha %>%
  ggplot(aes(x=disease_stat, y=invsimpson, fill=disease_stat)) +
  stat_summary(fun = median, show.legend=FALSE, geom="crossbar") +
  geom_jitter(show.legend=FALSE, width=0.25, shape=21, color="black") +
  labs(x=NULL,
       y="Inverse Simpson Index") +
  scale_x_discrete(breaks=c("NonDiarrhealControl","DiarrhealControl","Case"),
                   labels=c(glue("Healthy<br>(N={healthy_n})"),
                            glue("Diarrhea and<br>*C.difficile* negative<br>\\
                                 (N={diarrhea_n})"),
                            glue("Diarrhea and<br>*C.difficile* positive<br>\\
                                 (N={case_n})"))
                   ) +
  scale_fill_manual(name=NULL,
                     breaks=c("NonDiarrhealControl","DiarrhealControl","Case"),
                     labels=c("Healthy",
                              "Diarrhea and<br>*C.difficile* negative",
                              "Diarrhea and<br>*C.difficile* positive"),
                     values=c(healthy_color, diarrhea_color, case_color)) +
  theme_classic() +
  theme(axis.text.x = element_markdown())

strip_chart +
  geom_line(data=tibble(x=c(2, 3), y=c(23, 23)),
            aes(x=x, y=y),
            inherit.aes=FALSE) +
  geom_line(data=tibble(x=c(1, 2.5), y=c(33, 33)),
            aes(x=x, y=y),
            inherit.aes=FALSE) +
  geom_text(data=tibble(x=2.5, y=24),
            aes(x=x, y=y, label="n.s."),
            inherit.aes=FALSE) +
  geom_text(data=tibble(x=1.75, y=33.5),
            aes(x=x, y=y, label="*"), size=8,
            inherit.aes=FALSE)


ggsave("schubert_diversity.tiff", width=4.5, height=3.5)

#schubert_roc.R
library(tidyverse)
library(readxl)
library(glue)
library(ggtext)

metadata <- read_excel(path="raw_data/schubert.metadata.xlsx", na="NA") %>%
  mutate(disease_stat = factor(disease_stat,
                               levels=c("NonDiarrhealControl",
                                        "DiarrhealControl",
                                        "Case")
                               )
  )

alpha_diversity <- read_tsv("raw_data/schubert.groups.ave-std.summary") %>%
  filter(method == "ave") %>%
  select(-label, -method)

disease_invsimpson <- inner_join(metadata, alpha_diversity,
                             by=c('sample_id'='group')) %>%
	select(disease_stat, invsimpson)


get_roc_data <- function(negative, positive){

  disease_invsimpson %>%
    filter(disease_stat == negative | disease_stat == positive) %>%
    mutate(disease_stat = recode(disease_stat,
                                 "{negative}" := FALSE,
                                 "{positive}" := TRUE)) %>%
    mutate(sens_spec = map(invsimpson, get_sens_spec, .)) %>%
    unnest(sens_spec) %>%
    mutate(comparison = glue("{negative}_{positive}"))

}

get_sens_spec <- function(x, data){

  predicted <- x > data$invsimpson

  tp <- sum(predicted & data$disease_stat)
  tn <- sum(!predicted & !data$disease_stat)
  fp <- sum(predicted & !data$disease_stat)
  fn <- sum(!predicted & data$disease_stat)

  specificity <- tn / (tn + fp)
  sensitivity <- tp / (tp + fn)


  tibble("sensitivity" = sensitivity, "specificity"=specificity)
}

roc_data <- bind_rows(
    get_roc_data("NonDiarrhealControl", "DiarrhealControl"),
    get_roc_data("NonDiarrhealControl", "Case"),
    get_roc_data("DiarrhealControl", "Case")
  ) %>%
  arrange(invsimpson)

roc_data %>%
  mutate(comparison = factor(comparison,
                             c("NonDiarrhealControl_DiarrhealControl",
                               "NonDiarrhealControl_Case",
                               "DiarrhealControl_Case"),
                             levels=c("NonDiarrhealControl_DiarrhealControl",
                                      "NonDiarrhealControl_Case",
                                      "DiarrhealControl_Case"))) %>%
  ggplot(aes(x=1-specificity, sensitivity, color=comparison)) +
  geom_abline(slope=1, intercept=0, color="gray") +
  geom_line(size = 1, linejoin="round") +
  labs(x="1-Specificity",
       y="Sensitivity") +
  scale_color_manual(name=NULL,
                     breaks=c("NonDiarrhealControl_DiarrhealControl",
                              "NonDiarrhealControl_Case",
                              "DiarrhealControl_Case"),
                     values=c("orange", "purple", "black"),
                     labels=c("Healthy vs. Diarrhea<br>(*C. difficile* negative)",
                               "Healthy vs. Diarrhea<br>(*C. difficile* positive)",
                               "*C. difficile* negative vs.<br>*C. difficile* positive (diarrhea)")) +
  theme_classic() +
  theme(
    legend.text = element_markdown(),
    legend.key.height = unit(25, "pt"),
    legend.position = c(0.8, 0.2)
  )

ggsave("schubert_roc_curves.tiff", width=5, height=4)