Figures for Alexis' 2014-04-17 seminar.

alspac-enrichment-figure.r

library(ggplot2)
library(plyr)


# Load data ---------------------------------------------------------------

enrich.file <- "results/alspac/enrichment/enrichment-results-alspac-encode-dnase-h3k4me3.csv"
enrich.df <- read.csv(enrich.file, stringsAsFactors = FALSE)

# address stem-cell samples missing tissue annotation
stem.rows <- which(is.na(enrich.df$tissue) & grepl("stem", enrich.df$description))
enrich.df$tissue[stem.rows] <- "undifferentiated stem cell"


# Prep data for plot ------------------------------------------------------

# log(p) threshold used to select enrichment results
plot.thresh <- 3.5 

plot.df <- subset(enrich.df, count >= 3 & !is.na(tissue))

# summarize enrichment for each tissue
enrich.tissue <- ddply(plot.df, .(feature, threshold, tissue), summarize,
                       mean = mean(enrichment), 
                       median = median(enrichment))

# order tissue levels by summarized enrichment
tissue.levels <- subset(enrich.tissue, 
                        threshold == plot.thresh & feature == "DNaseI")
tissue.levels <- arrange(tissue.levels, median)$tissue

enrich.tissue$tissue <- factor(enrich.tissue$tissue, tissue.levels)
plot.df$tissue <- factor(plot.df$tissue, tissue.levels)


# Create plot -------------------------------------------------------------

ggplot(subset(plot.df, threshold == plot.thresh)) + 
  aes(tissue, enrichment, color = feature) + 
  stat_summary(fun.y = median, aes(group = feature), geom = "line") +
  stat_summary(fun.y = median, geom = "point") +
  geom_point(shape = 3, alpha = 0.6) +
  scale_color_brewer("Feature", palette = "Set1") +
  coord_flip() +
  ylab("Fold enrichment") +
  theme_bw() +
  theme(legend.position = "top", axis.title.y = element_blank())