Introduction to R (part 1)

Overview

Set Up

Materials

All prerequisites, links to material and slides for this course can be found on github.

Intro_To_R_1

Or can be downloaded as a zip archive from here.

Download zip

Course materials

Once the zip file in unarchived. All presentations as HTML slides and pages, their R code and HTML practical sheets will be available in the directories underneath.

r_course/presentations/slides/ Presentations as an HTML slide show.
r_course/presentations/singlepage/ Presentations as an HTML single page.
r_course/presentations/r_code/ R code in presentations.
r_course/exercises/ Practicals as HTML pages.
r_course/answers/ Practicals with answers as HTML pages and R code solutions.

Set the Working directory

Before running any of the code in the practicals or slides we need to set the working directory to the folder we unarchived.

You may navigate to the unarchived Intro_To_R_1Day folder in the Rstudio menu. We specifically want to set the working directory to the r_course folder.

Session -> Set Working Directory -> Choose Directory

or in the console.

setwd("/PathToMyDownload/Intro_To_R_1Day-master/r_course")
# e.g. setwd("~/Downloads/Intro_To_R_1Day/r_course")

Background of R

What is R?

R is a scripting language and environment for statistical computing.

Developed by Robert Gentleman and Ross Ihaka.

Inheriting much from S (Bell labs).

Suited to high level data analysis
Open source & cross platform
Extensive graphics capabilities
Diverse range of add-on packages
Active community of developers
Thorough documentation

What is R to you?

R comes with excellent “out-of-the-box” statistical and plotting capabilities.

R provides access to 1000s of packages (CRAN/MRAN/R-forge) which extend the basic functionality of R while maintaining high quality documentation.

In particular, Robert Gentleman developed the Bioconductor project where 100’s of packages are directly related to computational biology and analysis of associated high-throughput experiments.

R use over time

How to get R?

Freely available from R-project website.

RStudio provides an integrated development environment (IDE) which is freely available from RStudio site

We will be using RStudio and R already installed on your machines.

R website RStudio website

A quick tour of RStudio

Four main panels - Scripting panel - R interface - Environment and history - Files, directories and help

Let’s load RStudio and take a look

RStudio appearance

Data Types in R

Different Data types in R

Simple calculations
Variables
Vectors
Lists
Matrices
Data frames

Simple Calculations

At its most basic, R can be used as a simple calculator.

> 3+1

## [1] 4

> 2*2

## [1] 4

> sqrt(25)-1

## [1] 4

Using functions

The sqrt(25) demonstrates the use of functions in R. A function performs a complex operation on it’s arguments and returns the result.

In R, arguments are provided to a function within the parenthesis – ( ) – that follows the function name. So sqrt(ARGUMENT) will provide the square root of the value of ARGUMENT.

Other examples of functions include min(), sum(), max().

Note multiple arguments are separated by a comma.

min(2,4,6)

## [1] 2

sum(2,4,6)

## [1] 12

max(2,4,6)

## [1] 6

Using functions

R has many useful functions “built in” and ready to use as soon as R is loaded.

An incomplete, illustrative list can be seen here

In addition to R standard functions, additional functionality can be loaded into R using libraries. These include specialised tools for areas such as sequence alignment, read counting etc.

If you need to see how a function works try ? in front of the function name.

?sqrt

Lets run ?sqrt in RStudio and look at the help.

Using functions

Arguments have names and order

With functions such as min() and sqrt(), the arguments to be provided are obvious and the order of these arguments doesnt matter.

min(5,4,6)

## [1] 4

min(6,4,5)

## [1] 4

Many functions however have an order to their arguments. Try and look at the arguments for the dir() function using ?dir.

?dir

Using functions

Setting names for arguments

Often we know the names of arguments but not necessarily their order. In cases where we want to be sure we specify the right argument, we provide names for the arguments used.

dir()
dir(full.names=T)

This also means we don’t have to copy out all the defaults for arguments preceeding it.

dir(full.names=T)
# Is equivalent to...
dir(".",NULL,FALSE,T)

Variables

As with other programming languages and even graphical calculators, R makes use of variables.

A variable stores a value as a letter or word.

In R, we make use of the assignment operator <-

x <- 10

Now x holds the value of 10

## [1] 10

Altering variables

## [1] 10

Variables can be altered in place

x <- 20
x

## [1] 20

Using variables

Variables can be used just as the values they contain.

## [1] 20

x + sqrt(25)

## [1] 25

Variables can be used to create new variables

y <- x + sqrt(25)
y

## [1] 25

Vectors

In R the most basic variable or data type is a vector. A vector is an ordered collection of values. The x and y variables we have previously assigned are examples of a vector of length 1.

## [1] 20

length(x)

## [1] 1

Vectors

To create a multiple value vector we use the function c() to combine the supplied arguments into one vector.

x <- c(1,2,3,4,5,6,7,8,9,10)
x

##  [1]  1  2  3  4  5  6  7  8  9 10

length(x)

## [1] 10

Vectors

Vectors of continuous stretches of values can be created using a colon (:) as a shortcut.

y <- 6:10
y

## [1]  6  7  8  9 10

Other useful function to create stretchs of numeric vectors are seq() and rep(). The seq() function creates a sequence of numeric values from a specified start and end value, incrementing by a user defined amount. The rep() function repeats a variable a user-defined number of times.

seq(from=1,to=5,by=2)

## [1] 1 3 5

rep(c(1,5,10),3)

## [1]  1  5 10  1  5 10  1  5 10

Indexing

Square brackets [] identify the position within a vector (the index). These indices can be used to extract relevant values from vectors.

z <- seq(from=2,to=20,by=2) 
z

##  [1]  2  4  6  8 10 12 14 16 18 20

z[1]

## [1] 2

z[8]

## [1] 16

Indexing

Indices can be used to extract values from multiple positions within a vector.

z[c(1,6)]

## [1]  2 12

Negative indices can be used to extract all positions except that specified.

z[-5]

## [1]  2  4  6  8 12 14 16 18 20

Indexing and replacement

We can use indices to modify a specific position in a vector.

##  [1]  2  4  6  8 10 12 14 16 18 20

z[5] <- 1000
z

##  [1]    2    4    6    8 1000   12   14   16   18   20

Indexing and replacement

Indices can be specified using other vectors.

## [1]  6  7  8  9 10

z[y] <- 0
z

##  [1]    2    4    6    8 1000    0    0    0    0    0

Remember!

Square brackets [] for indexing.

x[1]

## [1] 1

Parentheses () for function argments.

sqrt(4)

## [1] 2

Arithmetic operations

Vectors in R can be used in arithmetic operations as seen with variables earlier. When a standard arithmetic operation is applied to vector, the operation is applied to each position in a vector.

x <- c(1,2,3,4,5,6,7,8,9,10)
x

##  [1]  1  2  3  4  5  6  7  8  9 10

y <- x*2
y

##  [1]  2  4  6  8 10 12 14 16 18 20

Multiple vectors can be used within arithmetic operations.

x+y

##  [1]  3  6  9 12 15 18 21 24 27 30

Arithmetic operations

When applying an arithmetic operation between two vectors of unequal length, the shorter will be recycled.

x <- c(1,2,3,4,5,6,7,8,9,10)
x

##  [1]  1  2  3  4  5  6  7  8  9 10

x+c(1,2)

##  [1]  2  4  4  6  6  8  8 10 10 12

x+c(1,2,3)

## Warning in x + c(1, 2, 3): longer object length is not a multiple of shorter
## object length

##  [1]  2  4  6  5  7  9  8 10 12 11

Character vectors

So far we have only looked at numeric vectors or variables.

In R we can also create character vectors again using c() function. These vectors can be indexed just the same.

y <- c("ICTEM","CommonWealth","Wolfson")
y[2]

## [1] "CommonWealth"

Character vectors can be used to assign names to other vectors.

x <- c(1:3)
names(x) <- y
x

##        ICTEM CommonWealth      Wolfson 
##            1            2            3

Character vectors as names

These named vectors maybe indexed by a position’s “name”.

x[c("ICTEM","Wolfson")]

##   ICTEM Wolfson 
##       1       3

Index names missing from vectors will return special value “NA”.

x[c("Strand")]

## <NA> 
##   NA

A note on NA values

In R, like many languages, when a value in a variable is missing, the value is assigned a NA value.

Similarly, when a calculation can not be perfomed, R will input a NaN value.

NA - Not Available.
NaN - Not A Number.

NA values allow for R to handle missing data correctly as they require different handling than standard numeric or character values. We will illustrate an example handling NA values later.

The unique() function

The unique() function can be used to retrieve all unique values from a vector.

geneList <- c("Gene1","Gene2","Gene3","Gene4","Gene5","Gene1","Gene3")
unique(geneList)

## [1] "Gene1" "Gene2" "Gene3" "Gene4" "Gene5"

Logical vectors

Logical vectors are a class of vector made up of TRUE or FALSE boolean values (single letter T/F can also be used).

z <-  c(TRUE,FALSE,TRUE,FALSE,TRUE,FALSE,TRUE,FALSE,TRUE,FALSE) 
# or
z <- c(T,F,T,F,T,F,T,F,T,F) 

z

##  [1]  TRUE FALSE  TRUE FALSE  TRUE FALSE  TRUE FALSE  TRUE FALSE

Logical vectors can be used like an index to specify postions in a vector. TRUE values will return the corresponding position in the vector being indexed.

x <- 1:10
x[z]

## [1] 1 3 5 7 9

The %in% operator

A common task in R is to subset one vector by the values in another vector.

The %in% operator in the context A %in% B creates a logical vector of whether values in A matches any values in of B.

This can be then used to subset the values within one character vector by a those in a second.

geneList <- c("Gene1","Gene2","Gene3","Gene4","Gene5","Gene1","Gene3")
secondGeneList <- c("Gene5","Gene3")
logical_index <- geneList %in% secondGeneList
logical_index

## [1] FALSE FALSE  TRUE FALSE  TRUE FALSE  TRUE

geneList[logical_index]

## [1] "Gene3" "Gene5" "Gene3"

Logical vectors from operators

Vectors may be evaluated to produce logical vectors. This can be very useful when using a logical to index.

Common examples are:

== evaluates as equal.
> and < evaluates as greater or less than respectively.
>= and <= evaluates as greater than or equal or less than or equal respectively.

x <- 1:10
x > 5

##  [1] FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE  TRUE  TRUE  TRUE

x[x > 5]

## [1]  6  7  8  9 10

Combining logical vectors

Logical vectors can be used in combination in order to index vectors. To combine logical vectors we can use some common R operators.

& - Requires both logical operators to be TRUE
| - Requires either logical operator to be TRUE.
! - Reverses the logical operator, so TRUE is FALSE and FALSE is TRUE.

x <- 1:10
!x > 4

##  [1]  TRUE  TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE

x > 4 & x < 7

##  [1] FALSE FALSE FALSE FALSE  TRUE  TRUE FALSE FALSE FALSE FALSE

x > 4 | x < 7

##  [1] TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE

Combining logical vectors

Such combinations can allow for complex selection of a vector’s values.

x <- 1:10
x

##  [1]  1  2  3  4  5  6  7  8  9 10

x[x > 4 & x < 7]

## [1] 5 6

x[x > 4 & !x < 7]

## [1]  7  8  9 10

Time for an exercise!

Exercise on vectors can be found here

Answers to exercise.

Answers can be found here here

Matrices

Creating matrices

In programs such as Excel we are used to tables.

igv

Creating matrices

All progamming languages have a concept of a table. In R, the most basic table type is a matrix.

A matrix can be created using the matrix() function with the arguments of nrow and ncol specifying the number of rows and columns respectively.

narrowMatrix <- matrix(1:10, nrow=5, ncol=2)
narrowMatrix

##      [,1] [,2]
## [1,]    1    6
## [2,]    2    7
## [3,]    3    8
## [4,]    4    9
## [5,]    5   10

wideMatrix <- matrix(1:10, nrow=2, ncol=5)
wideMatrix

##      [,1] [,2] [,3] [,4] [,5]
## [1,]    1    3    5    7    9
## [2,]    2    4    6    8   10

Creating matrices

By default when creating a matrix using the matrix function, the values fill the matrix by columns. To fill a matrix by rows the byrow argument must be set to TRUE.

wideMatrix <- matrix(1:10, nrow=2, ncol=5)
wideMatrix

##      [,1] [,2] [,3] [,4] [,5]
## [1,]    1    3    5    7    9
## [2,]    2    4    6    8   10

wideMatrixByRow <- matrix(1:10, nrow=2, ncol=5, byrow=TRUE)
wideMatrixByRow

##      [,1] [,2] [,3] [,4] [,5]
## [1,]    1    2    3    4    5
## [2,]    6    7    8    9   10

Finding dimensions

To find dimensions of a matrix, the dim() function will provide dimensions as the row then column number while nrow() and ncol() will return just row number and column number respectively.

dim(narrowMatrix)

## [1] 5 2

nrow(narrowMatrix)

## [1] 5

ncol(narrowMatrix)

## [1] 2

Joining vectors and matrices

A matrix can be created from multiple vectors or other matrices.

cbind() can be used to attach data to a matrix as columns.

x <- 1:5
y <- 11:15
z <- 21:22
newMatrix <- cbind(x,y)
newMatrix

##      x  y
## [1,] 1 11
## [2,] 2 12
## [3,] 3 13
## [4,] 4 14
## [5,] 5 15

Joining vectors and matrices

rbind() functions to bind to a matrix as rows.

newerMatrix <- rbind(newMatrix,z)
newerMatrix

##    x  y
##    1 11
##    2 12
##    3 13
##    4 14
##    5 15
## z 21 22

Joining vectors and matrices

Incompatible vectors and matrices

When creating a matrix using cbind() or matrix() from incompatible vectors then the shorter vector is recycled.

recycledMatrix2 <- matrix(1:5,ncol=2,nrow=3)

## Warning in matrix(1:5, ncol = 2, nrow = 3): data length [5] is not a
## sub-multiple or multiple of the number of rows [3]

recycledMatrix2

##      [,1] [,2]
## [1,]    1    4
## [2,]    2    5
## [3,]    3    1

Joining vectors and matrices

Incompatible vectors and matrices

For rbind() function, the longer vector is clipped.

recycledMatrix3 <- rbind(recycledMatrix2,c(1:5))

## Warning in rbind(...): number of columns of result is not a multiple of vector
## length (arg 2)

recycledMatrix3

##      [,1] [,2]
## [1,]    1    4
## [2,]    2    5
## [3,]    3    1
## [4,]    1    2

Column and row names

As we have seen with vectors, matrices can be named. For matrices the naming is done by columns and rows using colnames() and rownames() functions.

namedMatrix <- matrix(1:10,ncol=5,nrow=2)
colnames(namedMatrix) <- paste("Column",1:5,sep="_")
rownames(namedMatrix) <- paste("Row",1:2,sep="_")
namedMatrix

##       Column_1 Column_2 Column_3 Column_4 Column_5
## Row_1        1        3        5        7        9
## Row_2        2        4        6        8       10

Column and row names

Information on matrix names can also be retreived using the same functions.

colnames(namedMatrix)

## [1] "Column_1" "Column_2" "Column_3" "Column_4" "Column_5"

rownames(namedMatrix)

## [1] "Row_1" "Row_2"

Indexing

Selecting and replacing portions of a matrix can be done by indexing using square brackets [] much like for vectors.

When indexing matrices, two values may be provided within the square brackets separated by a comma to retrieve information on a matrix position.

The first value(s) corresponds to row(s) and the second to column(s).

myMatrix[rowOfInterest,columnOfInterest]

Indexing

narrowMatrix

##      [,1] [,2]
## [1,]    1    6
## [2,]    2    7
## [3,]    3    8
## [4,]    4    9
## [5,]    5   10

Value of first column, second row

narrowMatrix[2,1]

## [1] 2

Indexing

Similarly, whole rows or columns can be extracted. Single rows and columns will return a vector.

Values of second column (row index is empty!)

narrowMatrix[,2]

## [1]  6  7  8  9 10

Values of third row (column index is empty!)

narrowMatrix[3,]

## [1] 3 8

Indexing

When multiple columns or row indices are specified, a matrix is returned.

Values of second and third row (column index is empty!)

narrowMatrix[c(2,3),]

##      [,1] [,2]
## [1,]    2    7
## [2,]    3    8

Indexing by name

As with vectors, names can be used for indexing when present

colnames(narrowMatrix) <- paste("Column",1:2,sep="_")
rownames(narrowMatrix) <- paste("Row",1:5,sep="_")
narrowMatrix[,"Column_1"]

## Row_1 Row_2 Row_3 Row_4 Row_5 
##     1     2     3     4     5

narrowMatrix["Row_1",]

## Column_1 Column_2 
##        1        6

narrowMatrix["Row_1","Column_1"]

## [1] 1

Advanced indexing

As with vectors, matrices can be subset by logical vectors

narrowMatrix

##       Column_1 Column_2
## Row_1        1        6
## Row_2        2        7
## Row_3        3        8
## Row_4        4        9
## Row_5        5       10

narrowMatrix[,1]

## Row_1 Row_2 Row_3 Row_4 Row_5 
##     1     2     3     4     5

narrowMatrix[,1] < 5

## Row_1 Row_2 Row_3 Row_4 Row_5 
##  TRUE  TRUE  TRUE  TRUE FALSE

Advanced indexing

narrowMatrix[narrowMatrix[,1] < 5,]

##       Column_1 Column_2
## Row_1        1        6
## Row_2        2        7
## Row_3        3        8
## Row_4        4        9

Arithmetic operations

As with vectors, matrices can have arithmetic operations applied to cells, rows, columns or the whole matrix

narrowMatrix

##       Column_1 Column_2
## Row_1        1        6
## Row_2        2        7
## Row_3        3        8
## Row_4        4        9
## Row_5        5       10

narrowMatrix[1,1]+2

## [1] 3

narrowMatrix[1,]+2

## Column_1 Column_2 
##        3        8

Arithmetic operations

mean(narrowMatrix)

## [1] 5.5

Replacement

As with vectors, matrices can have their elements replaced

narrowMatrix

##       Column_1 Column_2
## Row_1        1        6
## Row_2        2        7
## Row_3        3        8
## Row_4        4        9
## Row_5        5       10

narrowMatrix[1,1] <- 10
narrowMatrix[,2] <- 1
narrowMatrix

##       Column_1 Column_2
## Row_1       10        1
## Row_2        2        1
## Row_3        3        1
## Row_4        4        1
## Row_5        5        1

Data types

Matrices must be all one type (i.e. numeric or character).

Here replacing one value with character will turn numeric matrix to character matrix.

narrowMatrix[,2] *2

## Row_1 Row_2 Row_3 Row_4 Row_5 
##     2     2     2     2     2

narrowMatrix[1,1] <- "Not_A_Number"
narrowMatrix

##       Column_1       Column_2
## Row_1 "Not_A_Number" "1"     
## Row_2 "2"            "1"     
## Row_3 "3"            "1"     
## Row_4 "4"            "1"     
## Row_5 "5"            "1"

Data types

narrowMatrix[,2] *2

## Error in narrowMatrix[, 2] * 2: non-numeric argument to binary operator

Time for an exercise!

Exercise on matrices can be found here

Answers to exercise.

Answers can be found here here

Factors

Creating factors

A special case of a vector is a factor.

Factors are used to store data which may be grouped in categories (categorical data). Specifying data as categorical allows R to properly handle the data and make use of functions specific to categorical data.

To create a factor from a vector we use the factor() function. Note that the factor now has an additional component called “levels” which identifies all categories within the vector.

vectorExample <- c("male","female","female","female")
factorExample <- factor(vectorExample)
factorExample

## [1] male   female female female
## Levels: female male

levels(factorExample)

## [1] "female" "male"

Summary() function

An example of the use of levels can be seen from applying the summary() function to the vector and factor examples

summary(vectorExample)

##    Length     Class      Mode 
##         4 character character

summary(factorExample)

## female   male 
##      3      1

Display order of levels

In our factor example, the levels have been displayed in an alphabetical order. To adjust the display order of levels in a factor, we can supply the desired display order to levels argument in the factor() function call.

factorExample <- factor(vectorExample, levels=c("male","female"))
factorExample

## [1] male   female female female
## Levels: male female

summary(factorExample)

##   male female 
##      1      3

Nominal factors

In some cases there is no natural order to the categories such that one category is greater than the other (nominal data). In this case we can see that R is gender neutral.

factorExample <- factor(vectorExample, levels=c("male","female"))
factorExample[1] < factorExample[2]

## Warning in Ops.factor(factorExample[1], factorExample[2]): '<' not meaningful
## for factors

## [1] NA

Ordinal factors

In other cases there will be a natural ordering to the categories (ordinal data). A factor can be specified to be ordered using the ordered argument in combination with specified levels argument.

factorExample <- factor(c("small","big","big","small"),
                        ordered=TRUE,levels=c("small","big"))
factorExample

## [1] small big   big   small
## Levels: small < big

factorExample[1] < factorExample[2]

## [1] TRUE

Replacement

Unlike vectors, replacing elements within a factor isn’t so easy. While replacing one element with an established level is possible, replacing with a novel element will result in a warning.

factorExample <- factor(c("small","big","big","small"))
factorExample[1] <- c("big")
factorExample

## [1] big   big   big   small
## Levels: big small

factorExample[1] <- c("huge")

## Warning in `[<-.factor`(`*tmp*`, 1, value = "huge"): invalid factor level, NA
## generated

factorExample

## [1] <NA>  big   big   small
## Levels: big small

Replacement

To add a new level we can use the levels argument.

levels(factorExample) <- c("big","small","huge")
factorExample[1] <- c("huge")
factorExample

## [1] huge  big   big   small
## Levels: big small huge

Data Frames

Creating data frames

We saw that with matrices you can only have one type of data. We tried to create a matrix with a character element and the entire matrix became a character.

In practice, we would want to have a table which is a mixture of types (e.g a table with sample names (character), sample type (factor) and survival time (numeric))

igv

Creating data frames

In R, we make use of the data frame object which allows us to store tables with columns of different data types. To create a data frame we can simply use the data.frame() function.

patientName <- c("patient1","patient2","patient3","patient4")
patientType <- factor(rep(c("male","female"),2))
survivalTime <- c(1,30,2,20)
dfExample <- data.frame(Name=patientName, Type=patientType, Survival_Time=survivalTime)
dfExample

##       Name   Type Survival_Time
## 1 patient1   male             1
## 2 patient2 female            30
## 3 patient3   male             2
## 4 patient4 female            20

Indexing and replacement

Data frames may be indexed just as matrices.

dfExample

##       Name   Type Survival_Time
## 1 patient1   male             1
## 2 patient2 female            30
## 3 patient3   male             2
## 4 patient4 female            20

dfExample[dfExample[,"Survival_Time"] > 10,]

##       Name   Type Survival_Time
## 2 patient2 female            30
## 4 patient4 female            20

Specify columns with $

Unlike matrices, it is possible to index a column by using the $ symbol.

dfExample <- data.frame(Name=patientName,Type=patientType,Survival_Time=survivalTime)
dfExample$Survival_Time

## [1]  1 30  2 20

dfExample[dfExample$Survival_Time < 10,]

##       Name Type Survival_Time
## 1 patient1 male             1
## 3 patient3 male             2

Specify columns with $

Using the $ allows for R to autocomplete your selection and so can speed up coding.

dfExample$Surv

## [1]  1 30  2 20

But this will not work..

dfExample[,"Surv"]

Create columns with $

The $ operator also allows for the creation of new columns for a data frame on the fly.

dfExample

##       Name   Type Survival_Time
## 1 patient1   male             1
## 2 patient2 female            30
## 3 patient3   male             2
## 4 patient4 female            20

dfExample$newColumn <- rep("newData",nrow(dfExample))
dfExample

##       Name   Type Survival_Time newColumn
## 1 patient1   male             1   newData
## 2 patient2 female            30   newData
## 3 patient3   male             2   newData
## 4 patient4 female            20   newData

Indexing and replacement

Replacement can occur in the same way we have seen for less complex data types: use the assignment operator <-

dfExample[dfExample[,"Survival_Time"] < 10,"Survival_Time"] <- 0
dfExample

##       Name   Type Survival_Time newColumn
## 1 patient1   male             0   newData
## 2 patient2 female            30   newData
## 3 patient3   male             0   newData
## 4 patient4 female            20   newData

Indexing and replacement

When we work with factors, for a replacement to be succesful it has to be a possible level.

dfExample[1,"Type"] <- "other"

## Warning in `[<-.factor`(`*tmp*`, iseq, value = "other"): invalid factor level,
## NA generated

dfExample

##       Name   Type Survival_Time newColumn
## 1 patient1   <NA>             0   newData
## 2 patient2 female            30   newData
## 3 patient3   male             0   newData
## 4 patient4 female            20   newData

Factors in data frames

It is possible to update factors in data frames just as with standard factors.

dfExample <- data.frame(Name=patientName,Type=patientType,
                        Survival_Time=survivalTime)

levels(dfExample[,"Type"]) <- c(levels(dfExample[,"Type"]) ,
                                "other")

dfExample[1,"Type"] <- "other"
dfExample

##       Name   Type Survival_Time
## 1 patient1  other             1
## 2 patient2 female            30
## 3 patient3   male             2
## 4 patient4 female            20

Factors in data frames

If you are running an older version of R (anything from before 4.0), then by default all columns with characters will be considered to be factors.

This is controlled when you make the data frame, by the stringsAsFactors argument. Up until recently the default parameter for this was TRUE. It now defaults to FALSE

dfExample <- data.frame(Name=patientName,
                        Type=patientType,
                        Survival_Time=survivalTime,
                        stringsAsFactors = T)


dfExample[dfExample[,"Survival_Time"] < 10,"Name"] <- "patientX"

## Warning in `[<-.factor`(`*tmp*`, iseq, value = c("patientX", "patientX")):
## invalid factor level, NA generated

dfExample

##       Name   Type Survival_Time
## 1     <NA>   male             1
## 2 patient2 female            30
## 3     <NA>   male             2
## 4 patient4 female            20

Factors in data frames

dfExample <- data.frame(Name=patientName,
                        Type=patientType,
                        Survival_Time=survivalTime,
                        stringsAsFactors = F)


dfExample[dfExample[,"Survival_Time"] < 10,"Name"] <- "patientX"
dfExample

##       Name   Type Survival_Time
## 1 patientX   male             1
## 2 patient2 female            30
## 3 patientX   male             2
## 4 patient4 female            20

Ordering with order()

A useful function in R is order()

For numeric vectors, order() by default returns the indices of a vector in that vector’s increasing order. This behaviour can be altered by using the “decreasing” argument passed to order.

testOrder <- c(20,10,30)
testOrder

## [1] 20 10 30

order(testOrder,decreasing=T)

## [1] 3 1 2

Ordering with order()

Once you have a vecor of ordered indices, you can then use this to index your object.

testOrder[order(testOrder)]

## [1] 10 20 30

testOrder[order(testOrder,decreasing=T)]

## [1] 30 20 10

Ordering with NA values

When a vector contains NA values, these NA values will, by default, be placed last in ordering indices. This can be controlled by na.last argument.

testOrder <- c(2,1,NA,3)
testOrder[order(testOrder,decreasing=T,na.last=T)]

## [1]  3  2  1 NA

testOrder[order(testOrder,decreasing=T,na.last=F)]

## [1] NA  3  2  1

Ordering data frames

Since the order argument returns an index of intended order for a vector, we can use the order() function to order data frames by certain columns.

dfExample

##       Name   Type Survival_Time
## 1 patientX   male             1
## 2 patient2 female            30
## 3 patientX   male             2
## 4 patient4 female            20

dfExample[order(dfExample$Surv, decreasing=T),]

##       Name   Type Survival_Time
## 2 patient2 female            30
## 4 patient4 female            20
## 3 patientX   male             2
## 1 patientX   male             1

Ordering data frames

We can also use order to arrange multiple columns in a data frame by providing multiple vectors to order() function. Ordering will be performed in order of arguments.

dfExample[order(dfExample$Type,
                dfExample$Survival,
                decreasing=T),]

##       Name   Type Survival_Time
## 3 patientX   male             2
## 1 patientX   male             1
## 2 patient2 female            30
## 4 patient4 female            20

Merging data frames

A common operation is to join two data frames by a column of common values.

dfExample <- data.frame(Name=patientName,Type=patientType,
                        Survival_Time=survivalTime)
dfExample

##       Name   Type Survival_Time
## 1 patient1   male             1
## 2 patient2 female            30
## 3 patient3   male             2
## 4 patient4 female            20

dfExample2 <- data.frame(Name=patientName[1:3],
                        height=c(6.1,5.1,5.5))
dfExample2

##       Name height
## 1 patient1    6.1
## 2 patient2    5.1
## 3 patient3    5.5

Merging data frames

To do this we can use the merge() function with the data frames as the first two arguments. We can then specify the columns to merge by with the by argument. To keep only data pertaining to values common to both data frames the all argument is set to FALSE.

mergedDF <- merge(dfExample,dfExample2,by=1,all=F)
mergedDF

##       Name   Type Survival_Time height
## 1 patient1   male             1    6.1
## 2 patient2 female            30    5.1
## 3 patient3   male             2    5.5

Time for an exercise!

Exercise on data frames can be found here

Answers to exercise.

Answers can be found here here

Lists

Creating lists

Lists are the final data type we will look at.

In R, lists provide a general container which may hold any data types of unequal lengths as part of its elements.

firstElement <- c(1,2,3,4)
secondElement <- matrix(1:10,nrow=2,ncol=5)
thirdElement <- data.frame(colOne=c(1,2,4,5),colTwo=c("One","Two","Three","Four"))

lists

To create a list we can simply use the list() function with arguments specifying the data we wish to include in the list.

myList <- list(firstElement,secondElement,thirdElement)
myList

## [[1]]
## [1] 1 2 3 4
## 
## [[2]]
##      [,1] [,2] [,3] [,4] [,5]
## [1,]    1    3    5    7    9
## [2,]    2    4    6    8   10
## 
## [[3]]
##   colOne colTwo
## 1      1    One
## 2      2    Two
## 3      4  Three
## 4      5   Four

Named lists

Just as with vectors, list elements can be assigned names.

myNamedList <- list(First=firstElement,Second=secondElement,
                    Third=thirdElement)
myNamedList

## $First
## [1] 1 2 3 4
## 
## $Second
##      [,1] [,2] [,3] [,4] [,5]
## [1,]    1    3    5    7    9
## [2,]    2    4    6    8   10
## 
## $Third
##   colOne colTwo
## 1      1    One
## 2      2    Two
## 3      4  Three
## 4      5   Four

Indexing

List, as with other data types in R can be indexed. In contrast to other types, using [] on a list will subset the list to another list of selected indices. To retrieve an element from a list in R , two square brackets [[]] must be used.

myList <- list(firstElement,secondElement,thirdElement)
myList[1]

## [[1]]
## [1] 1 2 3 4

myList[[1]]

## [1] 1 2 3 4

As with data.frames, the $ sign may be used to extract named elements from a list

myNamedList$First

## [1] 1 2 3 4

Joining lists

Again, similar to vectors, lists can be joined together in R using the c() function

myNamedList <- list(First=firstElement,Second=secondElement,
                    Third=thirdElement)
myNamedList <- c(myNamedList,list(fourth=c(4,4)))
myNamedList[c(1,4)]

## $First
## [1] 1 2 3 4
## 
## $fourth
## [1] 4 4

Joining vectors to lists

Note that on last slide we are joining two lists. If we joined a vector to a list, all elements of the vector would become list elements.

myList <- c(myList,c(4,4))
myList[3:5]

## [[1]]
##   colOne colTwo
## 1      1    One
## 2      2    Two
## 3      4  Three
## 4      5   Four
## 
## [[2]]
## [1] 4
## 
## [[3]]
## [1] 4

Flattening lists

Sometimes you will wish to “flatten” out a list. When a list contains compatable objects, i.e. list of all one type, the unlist() function can be used. Note the maintenance of names with their additional sufficies.

myNamedList <- list(First=c(1,2,3),Second=c(2,6,7),Third=c(1,4,7))
myNamedList

## $First
## [1] 1 2 3
## 
## $Second
## [1] 2 6 7
## 
## $Third
## [1] 1 4 7

flatList <- unlist(myNamedList)
flatList[1:7]

##  First1  First2  First3 Second1 Second2 Second3  Third1 
##       1       2       3       2       6       7       1

Flattening lists to matrices

A common step is to turn a list of standard results into matrix. This can be done in a few steps in R, using functions you are now familair with.

myNamedList <- list(First=c(1,2,3),Second=c(2,6,7),Third=c(1,4,7))
flatList <- unlist(myNamedList)
listAsMat <- matrix(flatList,
                    nrow=length(myNamedList),
                    ncol=3,
                    byrow=T,
                    dimnames=list(names(myNamedList)))
listAsMat

##        [,1] [,2] [,3]
## First     1    2    3
## Second    2    6    7
## Third     1    4    7

Coercing data formats

Identifying an objects class

We can use the class() function to tell us the class of our variable or object.

class(namedMatrix)

## [1] "matrix" "array"

class(dfExample)

## [1] "data.frame"

Coercing data class

Sometime you may want to cast an object into a new class There are a group of functions to coerce your data object to a new format, as._desiredClass_().

namedMatrix

##       Column_1 Column_2 Column_3 Column_4 Column_5
## Row_1        1        3        5        7        9
## Row_2        2        4        6        8       10

as.character(namedMatrix)

##  [1] "1"  "2"  "3"  "4"  "5"  "6"  "7"  "8"  "9"  "10"

as.vector(namedMatrix)

##  [1]  1  2  3  4  5  6  7  8  9 10

Coercing data class

as.data.frame(namedMatrix)

##       Column_1 Column_2 Column_3 Column_4 Column_5
## Row_1        1        3        5        7        9
## Row_2        2        4        6        8       10

as.list(namedMatrix)

## [[1]]
## [1] 1
## 
## [[2]]
## [1] 2
## 
## [[3]]
## [1] 3
## 
## [[4]]
## [1] 4
## 
## [[5]]
## [1] 5
## 
## [[6]]
## [1] 6
## 
## [[7]]
## [1] 7
## 
## [[8]]
## [1] 8
## 
## [[9]]
## [1] 9
## 
## [[10]]
## [1] 10

Reading and Writing Data

Data from external sources

Most of the time, you will not be generating data in R but will be importing data from external files.

A standard format for this data is a table:

Gene_Name	Sample_1.hi	Sample_2.hi	Sample_3.hi
Gene_a	3.475048	2.956164	3.419013
Gene_b	3.797961	2.817407	4.620163
Gene_c	4.053925	4.016550	3.936086
Gene_d	4.529539	4.619006	4.988511
Gene_e	10.057140	8.846590	10.553874
Gene_f	10.104833	11.810654	10.875768
Gene_g	10.148655	9.677445	9.047978
Gene_h	10.016013	8.428591	9.841227

Data from text file with read.table()

Tables from text files can be read with read.table() function

Table <- read.table("data/readThisTable.csv",sep=",",header=T)
Table[1:4,1:3]

##   Gene_Name Sample_1.hi Sample_2.hi
## 1    Gene_a    4.570237    3.230467
## 2    Gene_b    3.561733    3.632285
## 3    Gene_c    3.797274    2.874462
## 4    Gene_d    3.398242    4.415202

Here we have provided two arguments. - sep argument specifies how columns are separated in our text file. (“,” for .csv, “ for .tsv) - header argument specifies whether columns have headers.

Row names in read.table()

read.table() allows for significant control over reading files through its many arguments. Have a look at options by using ?read.table

The row.names argument can be used to specify a column to use as row names for the resulting data frame. Here we use the first column as row names.

Table <- read.table("data/readThisTable.csv",sep=",",header=T,row.names=1)
Table[1:4,1:3]

##        Sample_1.hi Sample_2.hi Sample_3.hi
## Gene_a    4.570237    3.230467    3.351827
## Gene_b    3.561733    3.632285    3.587523
## Gene_c    3.797274    2.874462    4.016916
## Gene_d    3.398242    4.415202    4.893561

Setting factors from read.table()

As mentioned, data which is read into R through read.table() will be of data frame class.

Similar to when we create data frames, we can control whether the data is read in as a factor or not with the stringsAsFactors argument. The default of this has also changed in R 4.0, from TRUE to FALSE.

Table <- read.table("data/readThisTable.csv", sep=",", header=T, stringsAsFactors=F)

Other very useful functions for read table include: - skip - To set number of lines to skip when reading. - comment.char - To set the start identifier for lines not to be read.

Data from other sources

The read.table function can also read data from http.

URL <- "http://rockefelleruniversity.github.io/readThisTable.csv"
Table <- read.table(URL,sep=",",header=T)
Table[1:2,1:3]

##   Gene_Name Sample_1.hi Sample_2.hi
## 1    Gene_a    4.111851    3.837018
## 2    Gene_b    6.047822    5.683518

And the clipboard (this is Windows version).

Table <- read.table(file="clipboard",sep=",",header=T)

Data from file columns

read.table() function will by default read every row and column of a file.

The scan() function allows for the selection of particular columns to be read into R and so can save memory when files are large.

x <- scan("data/readThisTable.csv",sep=",",
what = c("character",rep("numeric", 6)),skip=1)
x[1:3]

## [1] "Gene_a"           "4.57023720364456" "3.23046698308814"

Writing data to file

Once we have our data analysed in R, we will want to export it to a file.

The most common method is to use the write.table() function

write.table(Table, file="data/writeThisTable.csv", sep=",")

Since our data has column names but no row names, I will provide the arguments col.names and row.names to write.table()

write.table(Table, file="data/writeThisTable.csv", sep=",", row.names =F, col.names=T)

Reviewing your data

It is always important to know what your data is. Especially when you are reading it in for the first time. We have used indexing to get a taste of the data frames so far. But there are two functions to quickly check your data. head() and tail() return the first or last 6 lines by default.

head(Table)

##   Gene_Name Sample_1.hi Sample_2.hi Sample_3.hi Sample_4.low Sample_5.low
## 1    Gene_a    4.111851    3.837018    4.360628     3.752517     4.368069
## 2    Gene_b    6.047822    5.683518    4.315889     3.381136     3.630273
## 3    Gene_c    2.597068    3.316300    3.681509     4.886520     4.318289
## 4    Gene_d    6.009197    5.927419    2.244701     6.574108     8.288831
## 5    Gene_e   10.152509   10.218200   10.004835     2.251603     1.805168
## 6    Gene_f   11.107868    9.592153   10.263975     3.567560     2.496475
##   Sample_1.low
## 1     3.421009
## 2     5.560802
## 3     5.097783
## 4     6.857291
## 5     2.396295
## 6     3.587755

Reviewing your data

tail(Table)

##   Gene_Name Sample_1.hi Sample_2.hi Sample_3.hi Sample_4.low Sample_5.low
## 3    Gene_c    2.597068    3.316300    3.681509     4.886520     4.318289
## 4    Gene_d    6.009197    5.927419    2.244701     6.574108     8.288831
## 5    Gene_e   10.152509   10.218200   10.004835     2.251603     1.805168
## 6    Gene_f   11.107868    9.592153   10.263975     3.567560     2.496475
## 7    Gene_g    8.705787    8.949422    9.226990    10.051516     7.841664
## 8    Gene_h    9.239039    9.839734   10.027812    11.084444     9.316200
##   Sample_1.low
## 3     5.097783
## 4     6.857291
## 5     2.396295
## 6     3.587755
## 7     9.649869
## 8     8.742943

head(Table,3)

##   Gene_Name Sample_1.hi Sample_2.hi Sample_3.hi Sample_4.low Sample_5.low
## 1    Gene_a    4.111851    3.837018    4.360628     3.752517     4.368069
## 2    Gene_b    6.047822    5.683518    4.315889     3.381136     3.630273
## 3    Gene_c    2.597068    3.316300    3.681509     4.886520     4.318289
##   Sample_1.low
## 1     3.421009
## 2     5.560802
## 3     5.097783

The rio (R io) package

We may want to import from formats other than plain text.

We can make use of an R package (the rio package) which allows us to import and export data to mulitple formats.

Formats include:

XML.
Matlab, SAS, SPSS and minitab output formats.
Excel and OpenOffice formats.

The rio package

To make use of the rio package functionality we will need to install this package to our version of R.

We can do this by using the install.packages() function with the package we wish to install.

install.packages(PACKAGENAME)

install.packages("rio")

The rio package

Once we have installed a package, we will need to load it to make the functions available to us.

We can load a library by using the library() function with package we wish to install

library(PACKAGENAME)

library("rio")

The rio package

The main two functions in the rio package are the import and export functions.

We can use the import() function to read in our csv file. We simple specify our file as an argument to the import() function.

import(Filename)

Table <- import("data/readThisTable.csv")
Table[1:2,]

##   Gene_Name Sample_1.hi Sample_2.hi Sample_3.hi Sample_4.low Sample_5.low
## 1    Gene_a    4.570237    3.230467    3.351827     3.930877     4.098247
## 2    Gene_b    3.561733    3.632285    3.587523     4.185287     1.380976
##   Sample_1.low
## 1     4.418726
## 2     5.936990

The rio package

By default we will only retrieve the first sheet.

We can specify the sheet by name or number using the which argument.

Table <- import("data/readThisXLS.xls", 
                which=2)
Table <- import("data/readThisXLS.xls", 
                which="Metadata")
Table[1:2,]

##       Patient Condition   Treatment
## 1 Sample_1.hi         A           X
## 2 Sample_2.hi         A NoTreatment

The rio package

If we want to import all sheets, we can use the import_list.

This returns a list containing our two spreadsheets. The list has two elements named after the corresponding XLS sheet.

Table <- import_list("data/readThisXLS.xls")
names(Table)

## [1] "ExpressionScores" "Metadata"

The rio package

Since this is a list of data.frames, we can access our sheets using standard list accessors, $ and [[]].

## Table[["ExpressionScores"]][1:2,]
Table$ExpressionScores[1:2,]

##   Gene_Name Sample_1.hi Sample_2.hi Sample_3.hi Sample_4.low Sample_5.low
## 1    Gene_a    4.570237    3.230467    3.351827     3.930877     4.098247
## 2    Gene_b    3.561733    3.632285    3.587523     4.185287     1.380976
##   Sample_1.low
## 1     4.418726
## 2     5.936990

Table$Metadata[1:2,]

##       Patient Condition   Treatment
## 1 Sample_1.hi         A           X
## 2 Sample_2.hi         A NoTreatment

The rio package

We can export our data back to file using the export() function and specifying the name of the output file to the file argument. The export() function will guess the format required from the extension.

ExpressionScores <- Table$ExpressionScores
export(ExpressionScores,file = "data/writeThisXLSX.xlsx")

The rio package

We can export a list of data.frames to Excel’s xlsx format using the export() function. The names of list elements will be used to name sheets in xlsx file.

names(Table) <- c("expr","meta")
export(Table,file = "data/writeThisMultipleXLSX.xlsx")

Time for an exercise!

Exercise on reading and writing data can be found here

Answers to exercise

Answers can be found here

Introduction to R - Session 1

Bioinformatics Resource Center - Rockefeller University

http://rockefelleruniversity.github.io/Intro_To_R_1Day/

brc@rockefeller.edu

Introduction to R (part 1)

Overview

Set Up

Materials

Course materials

Set the Working directory

Background of R

What is R?

What is R to you?

How to get R?

A quick tour of RStudio

Data Types in R

Different Data types in R

Simple Calculations

Using functions

Using functions

Using functions

Using functions

Variables

Variables

Altering variables

Using variables

Vectors

Vectors

Vectors

Vectors

Indexing

Indexing

Indexing and replacement

Indexing and replacement

Remember!

Arithmetic operations

Arithmetic operations

Character vectors

Character vectors as names

A note on NA values

The unique() function

Logical vectors

The %in% operator

Logical vectors from operators

Combining logical vectors

Combining logical vectors

Time for an exercise!

Answers to exercise.

Matrices

Creating matrices

Creating matrices

Creating matrices

Finding dimensions

Joining vectors and matrices

Joining vectors and matrices

Joining vectors and matrices

Incompatible vectors and matrices

Joining vectors and matrices

Incompatible vectors and matrices

Column and row names

Column and row names

Indexing

Indexing

Indexing

Indexing

Indexing by name

Advanced indexing

Advanced indexing

Arithmetic operations

Arithmetic operations

Replacement

Data types

Data types

Time for an exercise!

Answers to exercise.

Factors

Creating factors

Summary() function

Display order of levels

Nominal factors