9.9 Exercise 5. Data frame manipulation

Create the script “exercise5.R” and save it to the “Rcourse/Module1” directory: you will save all the commands of exercise 5 in that script.
Remember you can comment the code using #.

correction

getwd()
setwd("Rcourse/Module1")
setwd("~/Rcourse/Module1")

9.9.1 Exercise 5a

1- Create the following data frame:

|43|181|M| |34|172|F| |22|189|M| |27|167|F|

With Row names: John, Jessica, Steve, Rachel.
And Column names: Age, Height, Sex.

correction

df <- data.frame(Age=c(43, 34, 22, 27), 
                 Height=c(181, 172, 189, 167),
                 Sex=c("M", "F", "M", "F"),
                 row.names = c("John", "Jessica", "Steve", "Rachel"),
                 stringsAsFactors=FALSE)

2- Check the structure of df with str().

correction

str(df)

3- Calculate the average age and height in df

Try different approaches: * Calculate the average for each column separately.

correction

mean(df$Age)
mean(df$Height)

• Calculate the average of both columns simultaneously using the apply() function.

correction

# we have to remove the Sex column: we can calculate the average only with numbers
apply(df[,-3], 2, mean)
apply(df[,1:2], 2, mean)
apply(df[,-grep("Sex", colnames(df))], 2, mean)

4- Add one row to df2: Georges who is 53 years old and 168 tall.

correction

# Georges= allows us to enter the row name at the same time as we add a row
df <- rbind(df, Georges=c(53, 168, "M"))

5- Change the row names of df so the data becomes anonymous: Use Patient1, Patient2, etc. instead of actual names.

correction

rownames(df) <- c("Patient1", "Patient2", "Patient3", "Patient4", "Patient5")
# try also the paste function!
rownames(df) <- paste("Patient", 1:5, sep="")

6- Create the data frame df2 that is a subset of df which will contain only the female entries.

correction

# which elements are female ("F" in the "Sex" colum)
df$Sex=="F"
# retrieve rows that contain the female entries, and save in df2
df2 <- df[df$Sex=="F",]

7- Create the data frame df3 that is a subset of df which will contain only entries of males taller than 170.

correction

# which entries are males
df$Sex=="M"
# which entries are greater than 170 in column "Height"
df$Sex=="M" & df$Height > 170
# retrieve rows that contain the males that are taller than 170, and save in df3
df3 <- df[df$Sex=="M" & df$Height > 170,]

9.9.2 Exercise 5b

1. Create two data frames mydf1 and mydf2 as:

mydf1:

|1|14| |2|12| |3|15| |4|10|

mydf2:

|1|paul| |2|helen| |3|emily| |4|john| |5|mark|

With column names: “id”, “age” for mydf1, and “id”, “name” for mydf2.

correction

mydf1 <- data.frame(id=1:4, age=c(14,12,15,10))
mydf2 <- data.frame(id=1:5, name=c("paul", "helen", "emily", "john", "mark"))

2- Merge mydf1 and mydf2 by their “id” column. Look for the help page of merge and/or Google it!

correction

# input 2 data frames
# "by" columns indicate by which column you want to merge the data
merge(x=mydf1, y=mydf2, by.x="id", by.y="id")
mydf3 <- merge(x=mydf1, y=mydf2, by="id")

3- Order mydf3 by decreasing age. Look for the help page of order.

correction

# order the age column (default is increasing order)
order(mydf3$age)
# order the age column by decreasing order
order(mydf3$age, decreasing = TRUE)
# order the whole data frame by the column age in decreasing order
mydf3[order(mydf3$age, decreasing = TRUE), ]

9.9.3 Exercise 5c

1- Using the download.file function, download this file to your current directory. (Right click on “this file” -> Copy link location to get the full path).

correction

# failing: download.file("https://github.com/sbcrg/CRG_RIntroduction/blob/master/genes_dataframe.RData", "genes_dataframe.RData")
download.file("https://public-docs.crg.es/biocore/sbonnin/Rcourse/genes_dataframe.RData", "genes_dataframe.RData")

2- The function dir() lists the files and directories present in the current directory: check if genes_dataframe.RData was copied.

correction

dir()

3- Load genes_dataframe.RData in your environment Use the load function.

correction

load("genes_dataframe.RData")

4- genes_dataframe.RData contains the df_genes object: is it now present in your environment?

correction

ls()

5- Explore df_genes and see what it contains You can use a variety of functions: str, head, tail, dim, colnames, rownames, class…

correction

str(df_genes)
head(df_genes)
tail(df_genes)
dim(df_genes)
colnames(df_genes)
rownames(df_genes)
class(df_genes)

6- Select rows for which pvalue_KOvsWT < 0.05 AND log2FoldChange_KOvsWT > 0.5. Store in the up object.

correction

# rows where pvalue_KOvsWT < 0.05 
df_genes$pvalue_KOvsWT < 0.05
# rows where log2FoldChange_KOvsWT > 0.5
df_genes$log2FoldChange_KOvsWT > 0.5
# rows that comply both of the above conditions
df_genes$pvalue_KOvsWT < 0.05 & df_genes$log2FoldChange_KOvsWT > 0.5
# select rows for which pvalue_KOvsWT < 0.05 AND log2FoldChange_KOvsWT > 0.5
up <- df_genes[df_genes$pvalue_KOvsWT < 0.05 & 
                 df_genes$log2FoldChange_KOvsWT > 0.5,]

How many rows (genes) were selected?

7- Select from the up object the Zinc finger protein coding genes (i.e. the gene symbol starts with Zfp). Use the grep() function.

correction

# extract gene symbol column
up$gene_symbol
# use grep to get the genes matching the pattern "Zfp"
up[grep("Zf", up$gene_symbol), ]

8- Select rows for which pvalue_KOvsWT < 0.05 AND log2FoldChange_KOvsWT is > 0.5 OR < -0.5. For the selection of log2FoldChange: give the abs function a try!
Store in the diff_genes object.

correction

# rows where pvalue_KOvsWT < 0.05
df_genes$pvalue_KOvsWT < 0.05
# rows where log2FoldChange_KOvsWT > 0.5
df_genes$log2FoldChange_KOvsWT > 0.5
# rows where log2FoldChange_KOvsWT < -0.5
df_genes$log2FoldChange_KOvsWT > -0.5
# rows where log2FoldChange_KOvsWT < -0.5 OR log2FoldChange_KOvsWT > 0.5
df_genes$log2FoldChange_KOvsWT > 0.5 | df_genes$log2FoldChange_KOvsWT > -0.5
# same as above but using the abs function
abs(df_genes$log2FoldChange_KOvsWT) > 0.5
# combine all required criteria
df_genes$pvalue_KOvsWT < 0.05 & abs(df_genes$log2FoldChange_KOvsWT) > 0.5
# extract corresponding entries
diff_genes <- df_genes[df_genes$pvalue_KOvsWT < 0.05 & 
                 abs(df_genes$log2FoldChange_KOvsWT) > 0.5,]

How many rows (genes) were selected?