Factors and Data frames

 

These exercises are about the factors and data frames sections of Introduction to R.

Exercise 1 - Factors

• Create a nominal factor called CellType containing: “DC1”,“DC1”,“DC1”,“NK”,“NK”,“Mono”,“Mono”,“DC2”,“NK”

CellType <- factor(c("DC1","DC1","DC1","NK","NK","Mono","Mono","DC2","NK"))
CellType

## [1] DC1  DC1  DC1  NK   NK   Mono Mono DC2  NK  
## Levels: DC1 DC2 Mono NK

• Modify the the third position of CellType to “Neu”, by modifying the levels of the factor.

levels(CellType) <- c(levels(CellType),"Neu")
CellType[3] <- "Neu"
CellType

## [1] DC1  DC1  Neu  NK   NK   Mono Mono DC2  NK  
## Levels: DC1 DC2 Mono NK Neu

• Create CellType2 with the same entries, but directly specify the levels to include: “DC1”, “DC2”, “Mono”, “NK”, “Neu”, “Bcell”, “Tcell”.

CellType2 <- factor(c("DC1","DC1","DC1","NK","NK","Mono","Mono","DC2","NK"), levels = c("DC1", "DC2", "Mono", "NK", "Neu", "Bcell", "Tcell"))
CellType2

## [1] DC1  DC1  DC1  NK   NK   Mono Mono DC2  NK  
## Levels: DC1 DC2 Mono NK Neu Bcell Tcell

• Use combine to increase the length of CellType2 to include: “Neu”,“Neu”,“Bcell”,“DC1”

CellType2 <- c(CellType2, factor(c("Neu","Neu","Bcell","DC1"), levels = c("DC1", "DC2", "Mono", "NK", "Neu", "Bcell", "Tcell")))
CellType2

##  [1] DC1   DC1   DC1   NK    NK    Mono  Mono  DC2   NK    Neu   Neu   Bcell
## [13] DC1  
## Levels: DC1 DC2 Mono NK Neu Bcell Tcell

• Summarize the number of entries for each cell type.

summary(CellType2)

##   DC1   DC2  Mono    NK   Neu Bcell Tcell 
##     4     1     2     3     2     1     0

• Reorder the summary to alphabetical order

levels(CellType2) <- c("Bcell","DC1", "DC2", "Mono", "Neu","NK","Tcell")
summary(CellType2)

## Bcell   DC1   DC2  Mono   Neu    NK Tcell 
##     4     1     2     3     2     1     0

• Create a ordinal factor named “Height” containing – high, low, mid, low, mid, low, mid, high, mid, high.

Height <- factor(c("high", "low", "mid", "low", "mid", "low", "mid", "high", "mid", "high"),ordered=T,levels=c("low", "mid", "high"))
Height

##  [1] high low  mid  low  mid  low  mid  high mid  high
## Levels: low < mid < high

• Using a logical index, create new factor of only those from “Height”” greater than low.

filteredHeight <- Height[Height > "low"]
filteredHeight

## [1] high mid  mid  mid  high mid  high
## Levels: low < mid < high

• Replace the last index in “Height” with veryHigh and create new factor with those greater than mid.

newFactor <- factor(Height,ordered=T,levels=c("low", "mid", "high","veryHigh"))
newFactor[length(newFactor)] <- "veryHigh"
newFactor[newFactor > "mid"]

## [1] high     high     veryHigh
## Levels: low < mid < high < veryHigh

Exercise 2 - Data frames

• Create data frame called Annotation with a column of gene names (“Gene_1”, “Gene_2”, “Gene_3”,“Gene_4”,“Gene_5”), ensembl gene names (“Ens001”, “Ens003”, “Ens006”, “Ens007”, “Ens010”), pathway information (“Glycolysis”, “TGFb”, “Glycolysis”, “TGFb”, “Glycolysis”) and gene lengths (100, 3000, 200, 1000,1200).

Annotation <- data.frame(geneNames=c("Gene_1", "Gene_2", "Gene_3","Gene_4","Gene_5"), ensembl=c("Ens001", "Ens003", "Ens006", "Ens007", "Ens010"),pathway=c("Glycolysis", "TGFb", "Glycolysis", "TGFb", "Glycolysis"),geneLengths=c(100, 3000, 200, 1000,1200))
Annotation

##   geneNames ensembl    pathway geneLengths
## 1    Gene_1  Ens001 Glycolysis         100
## 2    Gene_2  Ens003       TGFb        3000
## 3    Gene_3  Ens006 Glycolysis         200
## 4    Gene_4  Ens007       TGFb        1000
## 5    Gene_5  Ens010 Glycolysis        1200

• Filter Annotation to geneLengths that are greater than 500 and less than 2000. Use the dollar sign to extract column information.

Annotation[Annotation$geneLengths>500 & Annotation$geneLengths<2000,]

##   geneNames ensembl    pathway geneLengths
## 4    Gene_4  Ens007       TGFb        1000
## 5    Gene_5  Ens010 Glycolysis        1200

• Check the data types of each column. Update the pathway column to be a factor.

class(Annotation[,1])

## [1] "character"

class(Annotation[,2])

## [1] "character"

class(Annotation[,3])

## [1] "character"

class(Annotation[,4])

## [1] "numeric"

Annotation[,3] <- factor(Annotation[,3])
Annotation

##   geneNames ensembl    pathway geneLengths
## 1    Gene_1  Ens001 Glycolysis         100
## 2    Gene_2  Ens003       TGFb        3000
## 3    Gene_3  Ens006 Glycolysis         200
## 4    Gene_4  Ens007       TGFb        1000
## 5    Gene_5  Ens010 Glycolysis        1200

class(Annotation[,3])

## [1] "factor"

• Create data frame called Sample1 with ensembl gene names (“Ens001”, “Ens003”, “Ens006”, “Ens010”) and expression (1000, 3000, 10000,5000)

Sample1 <- data.frame(ensembl=c("Ens001", "Ens003", "Ens006","Ens010"),expression=c(1000, 3000, 10000,5000))
Sample1

##   ensembl expression
## 1  Ens001       1000
## 2  Ens003       3000
## 3  Ens006      10000
## 4  Ens010       5000

• Create data frame called Sample2 with ensembl gene names (“Ens001”, “Ens003”, “Ens006”, “Ens007”,“Ens010”) and expression (1500, 1500, 17000,500,10000)

Sample2 <- data.frame(ensembl=c("Ens001", "Ens003", "Ens006","Ens007","Ens010"),expression=c(1500, 1500, 17000,500,10000))
Sample2

##   ensembl expression
## 1  Ens001       1500
## 2  Ens003       1500
## 3  Ens006      17000
## 4  Ens007        500
## 5  Ens010      10000

• Create a data frame containing only those gene names common to all data frames with all information from Annotation and the expression from Sample 1 and Sample 2.

AnnoSample1 <- merge(Annotation,Sample1,by.x=2,by.y=1,all=F) 
AnnoSample1And2 <- merge(AnnoSample1,Sample2,by=1,all=F) 
AnnoSample1And2

##   ensembl geneNames    pathway geneLengths expression.x expression.y
## 1  Ens001    Gene_1 Glycolysis         100         1000         1500
## 2  Ens003    Gene_2       TGFb        3000         3000         1500
## 3  Ens006    Gene_3 Glycolysis         200        10000        17000
## 4  Ens010    Gene_5 Glycolysis        1200         5000        10000

• Order our new dataframe by geneLengths - biggest to smallest.

AnnoSample1And2 <- AnnoSample1And2[order(AnnoSample1And2$geneLengths, decreasing = T),]
AnnoSample1And2

##   ensembl geneNames    pathway geneLengths expression.x expression.y
## 2  Ens003    Gene_2       TGFb        3000         3000         1500
## 4  Ens010    Gene_5 Glycolysis        1200         5000        10000
## 3  Ens006    Gene_3 Glycolysis         200        10000        17000
## 1  Ens001    Gene_1 Glycolysis         100         1000         1500

• Add an extra two columns containing the length normalized expressions for Sample 1 and Sample 2

AnnoSample1And2$Sample1_lne <- AnnoSample1And2$expression.x/AnnoSample1And2$geneLengths
AnnoSample1And2$Sample2_lne <- AnnoSample1And2$expression.y/AnnoSample1And2$geneLengths
AnnoSample1And2

##   ensembl geneNames    pathway geneLengths expression.x expression.y
## 2  Ens003    Gene_2       TGFb        3000         3000         1500
## 4  Ens010    Gene_5 Glycolysis        1200         5000        10000
## 3  Ens006    Gene_3 Glycolysis         200        10000        17000
## 1  Ens001    Gene_1 Glycolysis         100         1000         1500
##   Sample1_lne Sample2_lne
## 2    1.000000    0.500000
## 4    4.166667    8.333333
## 3   50.000000   85.000000
## 1   10.000000   15.000000

• Identify the mean length normalized expression across Sample1 and Sample2 for Ens006 genes

rownames(AnnoSample1And2) <- AnnoSample1And2$ensembl
mean(c(AnnoSample1And2["Ens006","Sample1_lne"],AnnoSample1And2["Ens006","Sample2_lne"]))

## [1] 67.5

• For all genes, identify the log2 fold change in length normalized expression from Sample 1 to Sample 2.

log2FoldChange <- log2(AnnoSample1And2$Sample2_lne) - log2(AnnoSample1And2$Sample1_lne)
names(log2FoldChange) <- AnnoSample1And2$geneNames
log2FoldChange

##     Gene_2     Gene_5     Gene_3     Gene_1 
## -1.0000000  1.0000000  0.7655347  0.5849625

• Identify the total length of genes in Glycolysis pathway.

sum(AnnoSample1And2[AnnoSample1And2$pathway == "Glycolysis","geneLengths"])

## [1] 1500