Introduction to R - Exercises

 # Practice with R

## Creating vectors/factors and dataframes

1. We are performing RNA-Seq on cancer samples being treated with three different types of treatment (A, B, and P). You have 12 samples total, with 4 replicates per treatment. Write the R code you would use to construct your metadata table as described below.  

1. Create the vectors/factors for each column.
2. Put them together into a dataframe called `meta`. (Hint: you can type out each vector/factor, or if you want the process go faster try exploring the rep() function). 
3. Use the row.names() function to assign row names to the dataframe.

Your finished metadata table should have information for the variables "sex", "stage", "treatment", and "myc" levels: 

## Subsetting vectors/factors and dataframes

 2. Using the metadata table from question #1, write out the R code you would use to perform the following operations (questions DO NOT build upon each other):

a. return only the "treatment" and "sex" columns:

b. return the "treatment" values for samples 5, 7, 9, and 10:

c. use subset() to return all data for those samples receiving treatment "P":

d. use subset() to return only the "stage" and "treatment" data for those samples with "myc" > 5000:

e. remove the "treatment" column from the dataset:

f.  remove samples 7, 8 and 9 from the dataset:

g. keep only samples 1-6:

h. add a column called "pre_treatment" to the beginning of the dataframe with the values T, F, F, F, T, T, F, T, F, F, T, T (Hint: use cbind()): 

i. change the names of the columns to: "A", "B", "C", "D":

• Finally, we would like to extract only those rows from the meta dataset for replicate 2 from all conditions (KD.2, OE.2, IR.2). Let's do this in steps:

1. 1. Write the function you would use to determine the row names of the meta dataset: 
      
      
   2. Using the which() function, write the R code you would run to determine the location of the row name KD.2 in the meta dataset: 
      
      
   3. Using the which() function, write the R code you would use to determine the location of row names KD.2, OE.2, and IR.2 in the meta dataset (use the OR operator ( | ) to return multiple locations):
       
   4. Now, extract the rows from the meta dataset with row names KD.2, OE.2, and IR.2 using a single line of code using nested functions: 

# ggplot2 for pretty figures

5. This metadata table describes an experiment that you have setup for RNA-seq analysis, while the associated count matrix gives the normalized counts for each sample for every gene. Download the count matrix and metadata using the links provided.

Create the same plot as above using ggplot2. Follow the instructions below to build your plot. Write the code you used and provide the final image.

a. Read in the metadata file using: meta <- read.delim("Mov10_full_meta.txt", sep="\t", row.names=1)

b. Read in the count matrix file using: data <- read.delim("normalized_counts.txt", sep="\t", row.names=1)

c. Create a vector called expression that contains the normalized count values from the row in in normalized_counts that corresponds to MOV10. 

d. Check the class of this expression vector. You will need to convert this to a numeric vector using as.numeric(expression)

e. Bind that vector to your metadata data frame (meta)  and call the new data frame df. 

f. Initialize a  ggplot with your df as input.

g. Add the geom_jitter() geometric object with the required aesthetics which are x and y.

h. Color the points based on sampletype

i. Add the theme_bw() layer

j. Add the title “MOV10” to the plot

k. Change the x-axis label to be blank

l. Change the y-axis label to “Normalized counts”

m. Using theme() change the following properties of the plot:

i) remove the legend (Hint: use ?theme help and scroll down to legend.position)

ii) change the plot title size to 1.5x the default

iii) change the axis title to 1.5x the default size

iv) change the size of the axis text only on the y-axis to 1.25x the default size