Practice Session & Exercise

Introduction to R and Basic Programming Concepts

Author

Affiliation

Bioinformatics Core Facility CECAD

1 Practice by Sessions

1.1 Session 2

1.1.1 Exercise 1:

Import the CSV file using readr from data/ directory.

Task: Read rna_results.csv into an object called results. Then print results.

Solution 1.

library(readr)

results <- read_csv("rna_results.csv")

results

# A tibble: 10 × 5
   gene   condition log2FC   padj baseMean
   <chr>  <chr>      <dbl>  <dbl>    <dbl>
 1 Tp53   Drug        2.31 0.001      1250
 2 Myc    Drug        1.85 0.004       980
 3 Cdkn1a Drug        3.12 0.0002      750
 4 Bcl2   Drug       -1.42 0.021       620
 5 Gapdh  Control     0.12 0.85       3200
 6 Actb   Control    -0.08 0.91       2800
 7 Il6    Drug        2.75 0.008       450
 8 Sox2   Control     0.43 0.42        310
 9 Nanog  Drug       -2.1  0.013       275
10 Oct4   Drug        1.67 0.032       190

1.1.2 Exercise 2:

Select only the following columns:

• gene
• log2FC
• padj

Solution 2.

library(dplyr)

results |> 
  select(gene, log2FC, padj)

# A tibble: 10 × 3
   gene   log2FC   padj
   <chr>   <dbl>  <dbl>
 1 Tp53     2.31 0.001 
 2 Myc      1.85 0.004 
 3 Cdkn1a   3.12 0.0002
 4 Bcl2    -1.42 0.021 
 5 Gapdh    0.12 0.85  
 6 Actb    -0.08 0.91  
 7 Il6      2.75 0.008 
 8 Sox2     0.43 0.42  
 9 Nanog   -2.1  0.013 
10 Oct4     1.67 0.032 

1.1.3 Exercise 3:

Find genes that are:

• significant: padj < 0.05
• upregulated: log2FC > 0

Bonus question

How would you find significant downregulated genes?

Solution 3.

results |> 
  filter( 
    padj < 0.05, 
    log2FC > 0 
  )

# A tibble: 5 × 5
  gene   condition log2FC   padj baseMean
  <chr>  <chr>      <dbl>  <dbl>    <dbl>
1 Tp53   Drug        2.31 0.001      1250
2 Myc    Drug        1.85 0.004       980
3 Cdkn1a Drug        3.12 0.0002      750
4 Il6    Drug        2.75 0.008       450
5 Oct4   Drug        1.67 0.032       190

# Bonus question

results |> 
  filter( 
    padj < 0.05, 
    log2FC < 0 
  )

# A tibble: 2 × 5
  gene  condition log2FC  padj baseMean
  <chr> <chr>      <dbl> <dbl>    <dbl>
1 Bcl2  Drug       -1.42 0.021      620
2 Nanog Drug       -2.1  0.013      275

1.1.4 Exercise 4:

Create a new column called direction.

The column should contain:

• "Up" when log2FC > 0
• "Down" when log2FC < 0

Solution 4.

results |>
  mutate(
    direction = if_else(
      log2FC > 0,
      "Up",
      "Down"
    )
  )

# A tibble: 10 × 6
   gene   condition log2FC   padj baseMean direction
   <chr>  <chr>      <dbl>  <dbl>    <dbl> <chr>    
 1 Tp53   Drug        2.31 0.001      1250 Up       
 2 Myc    Drug        1.85 0.004       980 Up       
 3 Cdkn1a Drug        3.12 0.0002      750 Up       
 4 Bcl2   Drug       -1.42 0.021       620 Down     
 5 Gapdh  Control     0.12 0.85       3200 Up       
 6 Actb   Control    -0.08 0.91       2800 Down     
 7 Il6    Drug        2.75 0.008       450 Up       
 8 Sox2   Control     0.43 0.42        310 Up       
 9 Nanog  Drug       -2.1  0.013       275 Down     
10 Oct4   Drug        1.67 0.032       190 Up       

1.1.5 Exercise 5:

Create a new column called significant.

It should contain:

• "Yes" when padj < 0.05
• "No" otherwise.

Solution 5.

results |>
  mutate(
    significant = if_else(
      padj < 0.05,
      "Yes",
      "No"
    )
  )

# A tibble: 10 × 6
   gene   condition log2FC   padj baseMean significant
   <chr>  <chr>      <dbl>  <dbl>    <dbl> <chr>      
 1 Tp53   Drug        2.31 0.001      1250 Yes        
 2 Myc    Drug        1.85 0.004       980 Yes        
 3 Cdkn1a Drug        3.12 0.0002      750 Yes        
 4 Bcl2   Drug       -1.42 0.021       620 Yes        
 5 Gapdh  Control     0.12 0.85       3200 No         
 6 Actb   Control    -0.08 0.91       2800 No         
 7 Il6    Drug        2.75 0.008       450 Yes        
 8 Sox2   Control     0.43 0.42        310 No         
 9 Nanog  Drug       -2.1  0.013       275 Yes        
10 Oct4   Drug        1.67 0.032       190 Yes        

1.1.6 Exercise 6:

Now combine both calculations.

Then keep only significant genes and display:

• gene
• log2FC
• padj
• direction

Solution 6.

results |>
  mutate(
    direction = if_else(
      log2FC > 0,
      "Up",
      "Down"
    ), 
    significant = if_else(
      padj < 0.05,
      "Yes",
      "No"
    ) 
  ) |> 
  filter(significant == "Yes") |> 
  select( 
    gene, 
    log2FC, 
    padj, 
    direction 
  )

# A tibble: 7 × 4
  gene   log2FC   padj direction
  <chr>   <dbl>  <dbl> <chr>    
1 Tp53     2.31 0.001  Up       
2 Myc      1.85 0.004  Up       
3 Cdkn1a   3.12 0.0002 Up       
4 Bcl2    -1.42 0.021  Down     
5 Il6      2.75 0.008  Up       
6 Nanog   -2.1  0.013  Down     
7 Oct4     1.67 0.032  Up       

1.1.7 Exercise 7:

Arrange the results so that the genes with the smallest adjusted p-values appear first.

Solution 7.

results |> 
  arrange(padj)

# A tibble: 10 × 5
   gene   condition log2FC   padj baseMean
   <chr>  <chr>      <dbl>  <dbl>    <dbl>
 1 Cdkn1a Drug        3.12 0.0002      750
 2 Tp53   Drug        2.31 0.001      1250
 3 Myc    Drug        1.85 0.004       980
 4 Il6    Drug        2.75 0.008       450
 5 Nanog  Drug       -2.1  0.013       275
 6 Bcl2   Drug       -1.42 0.021       620
 7 Oct4   Drug        1.67 0.032       190
 8 Sox2   Control     0.43 0.42        310
 9 Gapdh  Control     0.12 0.85       3200
10 Actb   Control    -0.08 0.91       2800

1.1.8 Exercise 8:

Arrange genes according to the absolute size of their log2 fold change, regardless of direction.

Which function can you use?

Solution 8.

results |>
  arrange(desc(abs(log2FC)))

# A tibble: 10 × 5
   gene   condition log2FC   padj baseMean
   <chr>  <chr>      <dbl>  <dbl>    <dbl>
 1 Cdkn1a Drug        3.12 0.0002      750
 2 Il6    Drug        2.75 0.008       450
 3 Tp53   Drug        2.31 0.001      1250
 4 Nanog  Drug       -2.1  0.013       275
 5 Myc    Drug        1.85 0.004       980
 6 Oct4   Drug        1.67 0.032       190
 7 Bcl2   Drug       -1.42 0.021       620
 8 Sox2   Control     0.43 0.42        310
 9 Gapdh  Control     0.12 0.85       3200
10 Actb   Control    -0.08 0.91       2800

# This puts the genes with the largest changes, regardless of direction, first.

1.1.9 Exercise 9:

Calculate the mean log2FC among significant genes. Make use of summarise()

Solution 9.

results |> 
  filter(padj < 0.05) |> 
  summarise( 
    mean_log2FC = mean(log2FC) 
  )

# A tibble: 1 × 1
  mean_log2FC
        <dbl>
1        1.17

1.2 Exercise 10:

From the RNA-seq results, find significant genes, classify them as Up or Down, count how many genes are in each direction, and calculate the average log2FC for each direction.

Solution 10.

results |> 
  filter(padj < 0.05) |> 
  mutate( 
    direction = if_else( 
      log2FC > 0, 
      "Up", 
      "Down" 
      ) 
    ) |> 
    group_by(direction) |> 
    summarise( n_genes = n(), 
    mean_log2FC = mean(log2FC)
  )

# A tibble: 2 × 3
  direction n_genes mean_log2FC
  <chr>       <int>       <dbl>
1 Down            2       -1.76
2 Up              5        2.34