#Get the differential expressed values from the comparrison interested,

#extract the normalized values from the assay of vsd and plot them.

#Giving condition and group from your design table

#1. Extract the counts.

dt <- mat

#2. Perform pca

pca_dt <- prcomp(t(dt))

cat("PCA running...\n")

# Sys.sleep(0.2)

#3. Extract percentVar data.

percentVar_dt <- pca_dt$sdev^2/sum(pca_dt$sdev^2)

cat("Percents calculated...\n")

# Sys.sleep(0.2)

#4. Create the new dataframe to plot.

dt_f <- data.frame(PC1=pca_dt$x[,comp1],

PC2=pca_dt$x[,comp2],

color_pca=color_pca,

shape_pca=shape_pca,

label_pca= label_pca)

cat("Data frame built...\n")

# Sys.sleep(0.2)

#5. Plot it

cat("Plotting...\n")

# Sys.sleep(0.2)

print(save_plot)

require(ggplot2)

require(ggrepel)

if (save_plot== "no") {

pca_p <- ggplot(data = dt_f, aes_string(x = paste0("PC1"),

y = paste0("PC2"),

color = "color_pca",

shape= "shape_pca", label="label_pca")) +

geom_point(size = 3) +

geom_text_repel(size= 3, max.overlaps = 50,

box.padding = 1.5,point.padding = 0.5,force = 50)+

xlab(paste0("PC", comp1,": ",

round(percentVar_dt[comp1] * 100), "% variance")) +

ylab(paste0("PC",comp2,": ",

round(percentVar_dt[comp2] * 100), "% variance")) +

# coord_fixed()+

NULL

}

if (save_plot== "yes"){

png(filename =paste0(name_of_plot,".png"),res = 300,width = 2560,height = 1440)

cat("Saving plot as: ",paste0(name_of_plot,"...\n"))

pca_p <- ggplot(data = dt_f, aes_string(x = paste0("PC",comp1),

y = paste0("PC",comp2),

color = "color_pca",

shape= "shape_pca", label="label_pca")) +

geom_text_repel(size= 3, max.overlaps = 50,

box.padding = 1.5,

point.padding = 0.5,force = 50)+

geom_point(size = 3) +

xlab(paste0("PC", comp1,": ", round(percentVar_dt[comp1] * 100), "% variance")) +

ylab(paste0("PC",comp2,": ", round(percentVar_dt[comp2] * 100), "% variance")) +

# coord_fixed()+

NULL

print(pca_p)

dev.off()

}

# Sys.sleep(0.2)

cat("Done")

print(pca_p)

#return(pca_p)