Lyn16 / rna_wrangle

My assignment of week 6

rna_wrangle

My assignment of week 6

Files need to be downloaded

git clone https://github.com/tgrn510/RNASeqExample.git

1. Install packages and import the data

#install.packages("dplyr")
#install.packages("reshape")
#install.packages("ggplot2")
library("dplyr")
library("reshape")
library("ggplot2")
setwd("/Users/lyn/test/RNASeqExample/RNASeqExample")
samples<-read.csv("sample_info.csv", header = TRUE, sep = ",", quote = "\"", dec = ".", fill = TRUE, row.names = 1)
genes<-read.csv("expression_results.csv", header = TRUE, sep = ",", quote = "\"", dec = ".", fill = TRUE, row.names = 1)
plot(density(log2(genes$KITA_02[(genes$KITA_02>0)])))
# row.names = 1 definded that gene names in axis or in the first column#

2. Returns the density data and plot the results

#Create a density distribution of expression values within genes$KITA_02 on log scales#
d <- density(genes$KITA_02)
plot(d)
#Create a density distribution of PF_BASES values within the sample_info.csv file#
e <- density(samples$PF_BASES)
plot(e)

3.plot a scatterplot of two columns against one another

plot(log2(genes$KITA_01[(genes$KITA_01>10 |genes$KITA_03>10 )]),log2(genes$KITA_03[(genes$KITA_01>10 |genes$KITA_03>10)]))

4. heatmap to show how well the samples correlate using the cor function

corr<-cor(genes)
melt_corr<-melt(corr)
head(melt_corr)
colnames(melt_corr)<-c("Var1", "Var2", "value")

ggplot(melt_corr, aes(x = Var1, y = Var2)) +
 geom_raster(aes(fill = value)) +
 scale_fill_gradient2(low="green", mid="white", high="red", midpoint=0.5) + theme_classic()

5. The dendrogram

genes_transsample <- t(genes[c(rep(FALSE,19),TRUE), ])
clusters <- hclust(dist(genes_transsample))
plot(clusters)

#do annotation on dendrograph#
#install.packages("dendextend")
library("dendextend")
dend <-as.dendrogram(clusters)
dend <-rotate(dend, 1:93)
dend <-color_branches(dend, k=4)
par(cex=0.5)#reduces font#
plot(dend)

6. Create a 3D principal component analysis

setwd("/Users/lyn/test/RNASeqExample/RNASeqExample")
samples <- read.csv("sample_info.csv",header = TRUE, sep = ",",quote = "\"" ,dec = ",", fill = TRUE, row.names = 1)
genes <- read.csv('expression_results.csv',header = TRUE, sep = ",", quote = "\"", dec = ".", fill = TRUE, row.names = 1)
min(genes[genes>0])
genes.log<-log2(genes+8.05e-12)
genes.log.small <- genes.log[seq(1, nrow(genes.log), 20), ]
pca <- prcomp(genes.log.small,center = TRUE,scale. = TRUE)

std_dev <- pca$sdev
  pr_var <- std_dev^2
  pr_var[1:10]
  prop_varex <- pr_var/sum(pr_var)

library(plotly)
#make it a interactive chart#
pcadf<-data.frame(pca$rotation)

#make it a 3D version chart##
samples$uid<-rownames(samples)
pcadf$uid<-rownames(pcadf)
samples<-inner_join(samples,pcadf,by="uid")
plot_ly(samples, x = ~PC2, y = ~PC3, z = ~PC5, size=~reads,marker = list(symbol = 'circle', sizemode = 'diameter'), sizes = c(5, 25), color = ~Kit, colors = c('#BF382A', '#0C4B8E')) %>%
  add_markers() %>%
  layout(scene = list(xaxis = list(title = 'PC2'),
  yaxis = list(title = 'PC3'),
  zaxis = list(title = 'PC5')))