Basic R programming. Ana Teresa Maia DCBM / CBME

Transcription

1 Basic R programming Ana Teresa Maia DCBM / CBME

2 Today Sources! Introduction Documentation and help Packages R Studio Basics and Syntax Data Types vectors; lists; data.frames; matrices R Programming Basic 25/06/12 Slide 2 of 62

3 Where did I get this from??!!! Thomas Girke, Stephen Eglen, on.action?uri=info:doi/ /journal.pcbi s001 BeniltonCarvalho, Simon Tavaré, and many others teaching at the Mphilin Computational Biology at the University of Cambridge R Programming Basic 25/06/12 Slide 3 of 62

4 Today Sources! Introduction Documentation and help Packages R Studio Basics and Syntax Data Types vectors; lists; data.frames; matrices R Programming Basic 25/06/12 Slide 4 of 62

5 Introduction to R Look and Feel of the R Environment R Library Depositories Getting Around Basic Syntax Data Types and Subsetting Basic Calculations Reading and Writing External Data Some Great R Functions Graphics Utilities R Programming Basic 25/06/12 Slide 5 of 62

6 What is R? Why use R? Computer language and complete statistical package which allows the user to program and use tools developed by others Efficient functions and data structures for data analysis Powerful graphics Support group and widely used Free, open source Versatile R Programming Basic 25/06/12 Slide 6 of 62

7 History S language came from Bell Labs (Becker, Chambers and Wilks). Commercial version S-plus (1988). R emerged as a combination of S and Scheme: Ross Ihakaand Robert Gentleman (NZ). 1993: first announcement. 1995: 0.60 release, now under GPL. Mar 30 th 2012: release Stable, multi-platform. Major release typically Apr/Oct with fixes between. R-core now 20 people, key academics in field, including John Chambers. R Programming Basic 25/06/12 Slide 7 of 62

8 Today Sources! Introduction Documentation and help Packages R Studio Basics and Syntax Data Types vectors; lists; data.frames; matrices R Programming Basic 25/06/12 Slide 8 of 62

9 Documentation and on-line help Introductory Statistics with R (Springer, Dalgaard). A first course in statistical programming with R (CUP, Braun and Murdoch). Computational Genome Analysis: An Introduction (Springer, Deonier, Tavaréand Waterman). R programming for Bioinformatics (CRC Press,Gentleman). R-help mailing list. Bioinformatics and Computational Biology Solutions Using R and Bioconductor(John Verzani, 2004) UCR Manual (Thomas Girke) A beginners guide to R (Springer, Zuur, Ieno, Meesters) Alot of material placed in the MSc folder in Google Docs R Programming Basic 25/06/12 Slide 9 of 62

10 Package Directories CRAN(>3000 packages) general data analysis BioConductor(>500 packages) bioscience data analysis Omegahat(>30 packages) programming interfaces R Programming Basic 25/06/12 Slide 10 of 62

11 Today Sources! Introduction Documentation and help Packages R Studio Basics and Syntax Data Types vectors; lists; data.frames; matrices R Programming Basic 25/06/12 Slide 11 of 62

12 The R look! R Gui: OS X R Gui: Windows Command-line R: Linux/OS X R Programming Basic 25/06/12 Slide 12 of 62

13 RStudio: Alternative Working Environment for R Integrated development environment (IDE) that works well for beginners and developers R Programming Basic 25/06/12 Slide 13 of 62

14 Today Sources! Introduction Documentation and help Packages R Studio Basics and Syntax Data Types vectors; lists; data.frames; matrices R Programming Basic 25/06/12 Slide 14 of 62

15 Aim This course aims to teach R as a general-purpose programming language. Topics to be mastered in this course include: Interactive use of R. Basic data types: vector, matrix, list, data.frame, factor, character. Writing scripts. Graphical facilities. File input/output. Vectorization. Numerics issues. R Programming Basic 25/06/12 Slide 15 of 62

16 Startup/Closing Starting R The R GUI versions under Windows and Mac OS X can be opened by doubleclicking their icons. Alternatively, one can start it by typing R in a terminal (default under Linux). Startup/Closing Behavior The R environment is controlled by hidden files in the startup directory:.rdata,.rhistoryand.rprofile(optional). History means you can automatically save all commands you type Rdata saves everything in memory (can be large- be careful) Best to rename these using save.image(file= S01_GeneProjectMay2012.RData ) save(myvec, file= S01_GeneProjectMay2012.RData ) savehistory(file= S01_GeneProjectMay2012.Rhistory ) R Programming Basic 25/06/12 Slide 16 of 62

17 Startup/Closing Closing R (Note) ## Closing R > q() Save workspace image? [y/n/c]: When responding with y, then the entire R workspace will be written to the.rdatafile which can become very large. Often it is sufficient to just save an analysis protocol in an R source file. This way one can quickly regenerate all data sets and objects. R Programming Basic 25/06/12 Slide 17 of 62

18 Basic Location Create an object with the assignment operator <- (or = ) > object <-... List objects in current R session > ls() Return content of current working directory > dir() Return path of current working directory > getwd() [1]"/Users/atmaia" Change current working directory > setwd("/home/user") # change to a new folder named BioComp2012 R Programming Basic 25/06/12 Slide 18 of 62

19 Basic Location Create an object with the assignment operator <- (or = ) > object <-... List objects in current R session > ls() Return content of current working directory > dir() Return path of current working directory > getwd() [1]"/Users/atmaia" Change current working directory > setwd("/home/user") R Programming Basic 25/06/12 Slide 19 of 62

20 Basic Syntax General R command syntax > object <-function(arguments) > object <-object[arguments] Execute an R script > source("my script.r") Execute an R script from command-line $ R CMD BATCH my script.r $ R --slave < my script.r Save scripts S01_xxxDate.R, S02_xxxDate.R, etc where xxx is project name R Programming Basic 25/06/12 Slide 20 of 62

21 Basic Syntax Finding help >?function Load a library > library("my_library") Summary of all functions within a library > library(help="my_library") Load library manual (PDF file) > vignette() R Programming Basic 25/06/12 Slide 21 of 62

22 Example -Package Documentation Suppose we are interested in making a flow chart. After browsing the list of contributed packages on the R website we decide to use diagram. First download, install and load the package diagram > install.packages("diagram") > library(diagram) List of functions with brief descriptions in the diagram package > help(package=diagram) View the vignette > vignette("diagram") Look at the documentation for the function plotmat() >?plotmat R Programming Basic 25/06/12 Slide 22 of 62

23 Getting Data into R Simple Excel SpreadSheet data Simple table read.table() read.csv() scan() Some common data types Microarray SNP NGS R Programming Basic 25/06/12 Slide 23 of 62

24 Getting Data into R Reading Affymetrix Data library(affy) require(affy) # Alternative affybatch <- ReadAffy(celfile.path="[Location of your data]") eset<-justrma() R Programming Basic 25/06/12 Slide 24 of 62

25 Reading and Writing External Data Import Data into R > read.delim("mydata.xls", sep="\t") > Survival_TP53 <-read.table("/volumes/ /Survival.txt", sep="\t", header=t) Export Data from R to File > write.table(myframe, file="myfile.xls", sep="\t", quote=f) R Programming Basic 25/06/12 Slide 25 of 62

26 Basic Operators and Calculations Comparison operators: ==,! =, <, >, <=, >= Example: > 1==1 Logical operators: AND: &, OR:, NOT:! Example: > x <-1:10 >y <-10:1 >x>y & x>5 Calculations: Example: > x + y; sum(x); mean(x); sd(x); sqrt(x) > apply(iris[,1:3], 1, mean) R Programming Basic 25/06/12 Slide 26 of 62

27 options() The function options() sets several global options that affect how R computes and displays results. To look at the value of a single option, getoption ("option name") Options reset to the defaults with each new R session, even if you save the workspace. For example suppose we want to change the maximum number of digits printed from 7 (default) to 15. > defaults <- options() # Save all default options > getoption("digits") [1] 7 > pi [1] > options(digits=15) # Change to 15 digits > pi [1] > options(defaults) # Restore all default options > getoption("digits") [1] 7 R Programming Basic 25/06/12 Slide 27 of 62

28 options() The function options() sets several global options that affect how R computes and displays results. To look at the value of a single option, getoption ("option name") Options reset to the defaults with each new R session, even if you save the workspace. For example suppose we want to change the maximum number of digits printed from 7 (default) to 15. > defaults <- options() # Save all default options > getoption("digits") [1] 7 > pi [1] > options(digits=15) # Change to 15 digits > pi [1] > options(defaults) # Restore all default options > getoption("digits") [1] 7 R Programming Basic 25/06/12 Slide 28 of 62

29 Today Sources! Introduction Documentation and help Packages R Studio Basics and Syntax Data Types vectors; lists; data.frames; matrices R Programming Basic 25/06/12 Slide 29 of 62

30 Data Types A vector contains an indexed set of values that are all of the same type: Logical (ex. TRUE, FALSE) Numeric (ex. 1,2,3 ) Complex (ex. 1,b,3) Character (ex. "a","b","c") The numeric type can be further broken into integer, single and double types R Programming Basic 25/06/12 Slide 30 of 62

31 Data Structures vector elements of the same type factor categorical list can contain objects of different types matrix table of numbers data.frame table of numbers and/or characters environment- hashtable functional R Programming Basic 25/06/12 Slide 31 of 62

32 Data Structures There is no need to declare the types of the variables > x <- c(1,2,3,4,5,6) > class(x) [1] "numeric" y<-10 > class(y) [1] "numeric" > z<- "a string" > class(z) [1] "character R Programming Basic 25/06/12 Slide 32 of 62

33 Data Structures There is no need to declare the types of the variables > x <- c(1,2,3,4,5,6) > class(x) [1] "numeric" y<-10 > class(y) [1] "numeric" > z<- "a string" > class(z) [1] "character R Programming Basic 25/06/12 Slide 33 of 62

34 > t<- data.frame(type=c(rep("case",2),rep("control",3),time=rnorm(5))) > t type 1 case 2 case 3 control 4 control 5 control > class(t) [1] "data.frame R Programming Basic 25/06/12 Slide 34 of 62

35 > t<- data.frame(type=c(rep("case",2),rep("control",3),time=rnorm(5))) > t type 1 case 2 case 3 control 4 control 5 control > class(t) [1] "data.frame R Programming Basic 25/06/12 Slide 35 of 62

36 Creating Vectors Two symbols can be used for assignment <-and = > x <- c(1,2,3,4,5,6) > x [1] > s="a flower" > s [1] "a flower" > a <- seq.int(5) ## fast for integers > a [1] R Programming Basic 25/06/12 Slide 36 of 62

37 Creating Vectors Two symbols can be used for assignment <-and = > x <- c(1,2,3,4,5,6) > x [1] > s="a flower" > s [1] "a flower" > a <- seq.int(5) ## fast for integers > a [1] R Programming Basic 25/06/12 Slide 37 of 62

38 Functions for Creating Vectors c concatenate :-integer sequence seq general sequence rep repetitive patterns vector vector of given length with default value R Programming Basic 25/06/12 Slide 38 of 62

39 Functions for Creating Vectors > seq(1,6) [1] > seq(from=100, by=1, length=5) [1] > 1:6 [1] > rep(1:2,3) [1] > vector(mode="character", length=5) [1] "" "" "" "" "" > vector(mode="logical", length=5) [1] FALSE FALSE FALSE FALSE FALSE R Programming Basic 25/06/12 Slide 39 of 62

40 Functions for Creating Vectors > seq(1,6) [1] > seq(from=100, by=1, length=5) [1] > 1:6 [1] > rep(1:2,3) [1] > vector(mode="character", length=5) [1] "" "" "" "" "" > vector(mode="logical", length=5) [1] FALSE FALSE FALSE FALSE FALSE R Programming Basic 25/06/12 Slide 40 of 62

41 Functions for Creating Vectors > good <- seq(from=100, by=1, length=5) [1] >bad<-(1:3) > good[-bad] ## exclude elements [1] R Programming Basic 25/06/12 Slide 41 of 62

42 Functions for Creating Vectors > good <- seq(from=100, by=1, length=5) [1] >bad<-(1:3) > good[-bad] ## exclude elements [1] R Programming Basic 25/06/12 Slide 42 of 62

43 VectorizedArithmetic Most arithmetic operations in the R language are vectorized, i.e. the operation is applied element-wise > 1:3+10:12 [1] > 1:3 [1] When one operand is shorter than the other it is recycled > rep(1:2, 3) [1] R Programming Basic 25/06/12 Slide 43 of 62

44 VectorizedArithmetic Most arithmetic operations in the R language are vectorized, i.e. the operation is applied element-wise > 1:3+10:12 [1] > 1:3 [1] When one operand is shorter than the other it is recycled > rep(1:2, 3) [1] R Programming Basic 25/06/12 Slide 44 of 62

45 Naming indexes of a vector > joe< c(24, 1.70) > joe > names( joe) > names( joe) < c( age, height ) > joe > joe[ height ] == joe[2] Referingto index by name rather than by position can make code more readable, and flexible. Cannot do things like x [1:4] easily though, since you need to name all four elements you want. Note: in second use of names()above, we are actually using the replacement function names<, see later. R Programming Basic 25/06/12 Slide 45 of 62

46 Common functions for vectors length() rev() sum(), cumsum(), prod(), cumprod() mean(), sd(), var(), median() min(), max(), range(), summary() exp(), log(), sin(), cos(), tan() [radians, not degrees] round(), ceil(), floor(), signif() sort(), order(), rank() which(), which.max() any(), all() R Programming Basic 25/06/12 Slide 46 of 62

47 Common functions for vectors Functions can be called within function calls; the following are equivalent: x < c(3, 2, 9, 4) y < exp(x); z1 < which(y > 20) ## case 1 z2 < which ( exp(x) > 20) ## case 2 all.equal(z1, z2) R Programming Basic 25/06/12 Slide 47 of 62

48 Matrices and Arrays Can be created using matrix and array Vectors with dimension attribute > x<-matrix(1:10, nrow=2) > dim(x) [1] 2 5 > x [,1] [,2] [,3] [,4] [,5] [1,] [2,] > as.vector(x) [1] R Programming Basic 25/06/12 Slide 48 of 62

49 Matrices and Arrays Can be created using matrix and array Vectors with dimension attribute > x<-matrix(1:10, nrow=2) > dim(x) [1] 2 5 > x [,1] [,2] [,3] [,4] [,5] [1,] [2,] > as.vector(x) [1] R Programming Basic 25/06/12 Slide 49 of 62

50 Lists Ordered set of elements that can be arbitrary R objects (vectors, functions, etc) Can be heterogeneous > lst = list(a=1:3, b="olá", c=sqrt) > lst $a [1] $b [1] "olá" $c function (x).primitive("sqrt") > lst$c(49) [1] 7 R Programming Basic 25/06/12 Slide 50 of 62

51 Lists Ordered set of elements that can be arbitrary R objects (vectors, functions, etc) Can be heterogeneous > lst = list(a=1:3, b="olá", c=sqrt) > lst $a [1] $b [1] "olá" $c function (x).primitive("sqrt") > lst$c(49) [1] 7 R Programming Basic 25/06/12 Slide 51 of 62

52 Environments Differ from list because they have no order All objects are stored by name Names must match exactly > e1=new.env() > e1[["a"]] <- 1:3 > assign("b", "olá", e1) > ls(e1) [1] "a" "b" R Programming Basic 25/06/12 Slide 52 of 62

53 Data Frames Special structure in R Used to hold a set of spreadsheetlike tables data.frame, the observations are the rows and the covariates the columns Can be treated like matrices Columns are vectors, but different columns can be vectors of different types Are in fact lists, and list subsettingcan also be used on them R Programming Basic 25/06/12 Slide 53 of 62

54 Data Frames > df <-data.frame(type=rep(c("case","control"),c(2,3)),time=rexp(5)) > df type time 1 case case control control control > df$time [1] R Programming Basic 25/06/12 Slide 54 of 62

55 Data Frames > df <-data.frame(type=rep(c("case","control"),c(2,3)),time=rexp(5)) > df type time 1 case case control control control > df$time [1] R Programming Basic 25/06/12 Slide 55 of 62

56 General SubsettingRules Subsettingby indices > myvec<-1:26; names(myvec) <-LETTERS > myvec[1:4] Subsettingby same length logical vectors > mylog<-myvec> 10 > myvec[mylog] Subsettingby field names > myvec[c("b", "K","M")] Special case > iris$species R Programming Basic 25/06/12 Slide 56 of 62

57 Some Great R Functions The unique() function to make vector entries unique > unique(iris$sepal.length); length(unique(iris$sepal.length)) The table()function counts the occurrences of entries > table(iris$species) The aggregate()function computes statistics of data aggregates > aggregate(iris[,1:4], by=list(iris$species), FUN=mean, na.rm=t) The %in% function returns the intersect between two vectors > month.name[month.name%in% c("may", "July")] The merge()function joins data frames based on a common key column > merge(frame1, frame2, by.x=1, by.y=1, all = TRUE) R Programming Basic 25/06/12 Slide 57 of 62

58 Bugs Common Bugs and Fixes Syntax Error Trailing + Error When Performing Operations R Programming Basic 25/06/12 Slide 58 of 62

59 Error: syntax error Possible causes: Incorrect spelling (of the function, variable, etc.) Including a "+" when copying code from the Console Having an extra parenthesis at the end of a function Having an extra bracket when subsetting R Programming Basic 25/06/12 Slide 59 of 62

60 Trailing + Possible causes: Not closing a function call with a parenthesis Not closing brackets when subsetting Not closing a function you wrote with a squiggly brace R Programming Basic 25/06/12 Slide 60 of 62

61 Error in...: requires numeric matrix/vector arguments Possible causes: 1. Objects are data frames, not matrices 2. Elements of the vectors are characters Possible solutions: 1. Coerce (a copy of) the data set to be a matrix, with the as.matrix() command 2. Coerce (a copy of) the vector to have numeric entries, with the as.numeric() command R Programming Basic 25/06/12 Slide 61 of 62

62 Example -Sample Script # Calculate the mean of x x = c(0,5,7,9,1,2,8) mean(x) # How and how not to wrap expressions long.variable.name <- 5 really.long.variable.name <- 7 # R views the first line as a complete expression. Thus the code is # treated as two separate expressions instead of one long expression. long.answer.name <- 500*factorial(long.variable.name) + sqrt(really.long.variable.name) # Here the first line is not a complete expression (trailing + sign) so # R continues reading lines of code until the expression is complete long.answer.name <- 500*factorial(long.variable.name) + sqrt(really.long.variable.name) # Writing two expressions on the same line requires a ; mean(x); var(x) R Programming Basic 25/06/12 Slide 62 of 62