STAT115 STAT225 BIST512 BIO298 - Intro to Computational Biology Yang Li Lin Liu Jan 29, 2014 1.

STAT115 STAT225 BIST512 BIO298 - Intro to Computational Biology

Yang LiLin Liu

Jan 29, 2014

1


• Unix part slides courtesy: John Brunelle

• You can check out more details in:– https://software.rc.fas.harvard.edu/

training/intro_unix/latest/#(1)

2


Sign up on Odyssey

• Very simple, just go to http://rc.fas.harvard.edu/, then click on Account and Access Request Forms (right top of the website on Quick Links section), then click on RC Account form, and then fill it in as below – we will take care of the rest!

3

STAT115 STAT225 BIST512 BIO298 - Intro to Computational Biology 4


Basic Unix Command

• Log in:• ssh

[email protected]

5


Basic Unix Command

• Upload or download files:Upload:

scp username@host dir/targetfilenameDownload:

scp dir/yourfilename username@host

6


CaSe SeNsItIvE

• In shell commands, abc will be different from ABC

7


Terminology and notation

• Folders are usually referred to as directories

• Locations in the filesystem, like /n/home00/cfest350, are called paths

• The directory and file names that make up a path are always separated by a forward-slashes

• The top of the hierarchy is /, ie the root directory

8


Terminology and notation

9


Navigating the system: ls

10


Download and unzip files

• wget https://software.rc.fas.harvard.edu/training/examples.tar.gz

• tar xvf examples.tar.gz

11


What if you get confused

• man ls– Use the arrow keys, page up/down keys,

or the SPACE to navigate– To search for a phrase of text, for

example the word time, type /time and hit ENTER• Hit n to go to the next occurrence• Hit N to go to the previous occurrence• Hit q to quit

12


Kill process

• top• kill• killall• Ctrl-c• Exercise: Run the command

~/examples/bin/ticktock, and kill it once you've had enough

13


Copy files• mkdir workshop• cd workshop• cp ~/examples/aaa .• cp ~/examples/bbb ~/examples/ccc .• cp aaa zzz• rsync: replacement for cp, but can be used

to copy files to/from remote computers– e.g. rsync -avz --progress mywork

username@hostname:~/mywork

14


Moving and removing

15


File permissions

• The -rw-r--r-- displays the file mode bits– The first character is the type (- for files,

d for directories, and other letters (b, c, l, s etc.) for special files

– Following that are three groups of three characters, for read, write, and execute permissions for user, group, and others

16

STAT115 STAT225 BIST512 BIO298 - Intro to Computational Biology 17


• r = 4, w = 2, x = 1, rwx = 7• chmod 755

18


Hidden files

19


File manipulation

• cat ~/examples/gpl-3.0.txt• less ~/examples/gpl-3.0.txt• File editors: vim/emacs/nano

20


More shell commands

21


Piping your commands

• cat ~/examples/answers.out | awk '{print $3}'

22


Exercises

• List the last 5 files in /bin by combining the ls and tail commands with a pipe

• Count the number of lines that contain the word free in ~/examples/gpl-3.0.txt

23


The shell environment

• echo $PATH• Change $PATH:• PATH=$PATH\:/dir/path ; export PATH

24


Submit a job

• bsub < yourscript.bsub• yourscript.bsub:

#!/bin/sh#BSUB -u linliu@harvard#BSUB -J hellwo_world#BSUB -o hellow_world.out#BSUB -e hellow_world.err#BSUB -q short_serialpython hellow_world.py > hellow_world.out

25


Load modules

• module load dir/software– http://oldrcwebsite.rc.fas.harvard.edu/

faq/modulelist

26


Final tips

• Google is extremely helpful if you want to write some shell scripts

27


Getting started

Where the scripts/commands are executed

Where plots/help displayed, and packages installed.

Where the CODE is scripted

Show the variables/functions in memory


Workspace Management

• Before jumping into R, it is important to ask ourselvesWhere am I?

>getwd()

–I want to be there…• setwd(“C://”)

–With who am I?• dir() # lists all the files in the working directory

–With who I can count on?• ls() #lists all the variables on the current session


Workplace Management (2)

Saving>save(x,file=“name.RData”) #Saves specific

objects>save.image(“name.Rdata”) #Saves the whole

workspace

Loading>load(“name.Rdata”)

‘?function’ and ‘??function’>? To get the documentation of the function>?? Find related functions to the query


R Objects• Almost all things in R are OBJECTS!

– Functions, datasets, results, etc… (graphs NO)

• OBJECTS are classified by two criteria– MODE: How objects are stored in R

• Character, numeric, logical, factor, list, function…• To obtain the mode of an object

> mode(object)

– CLASS: How objects are treated by functions• Vector, matrix, array, data.frame,…• To obtain the class of an object

> class(object)


R classescharacter

> assembly = “hg19”> assembly> class(assembly)

numeric> expression = 3.456> expression> class(expression)

integer> nbases = “3000000000L”> nbases> class(nbases)

logical> completed = FALSE> completed> class(completed)


R classes - Vectorvector

>x=c(10,5,3,6); x[3:4]; x[1]

Computations on vector are performed on each entry of the vector

>y=c(log(x),x,x^2)

Not necessarily to have vectors of the same length in operations!

>w=sqrt(x)+2>z=c(pi,exp(1),sqrt(2))>x+z

–Logical vectors>aux=x<7


R Classes - Listlist

A vector of values of possibly different classes and different length.

Creating it.>x1 = 1:5>x2 = c(T,T,F,T,F)>y=list(question.number = x1, question.answer = x2)

Accesing it.>y;class(y)>y$question.answer[3]; y[[2]][3];

y[[“question.answer”]][3]>y$question.number[which(question.answer == T)]


R classes - Matrixmatrix

>x=1:8>dim(x)=c(2,4)>y=matrix(1:8,2,4,byrow=F)

Operations are applied on each element

>x*x; max(x)>x=matrix(1:28,ncol=4);

y=7:10 so then x*y is…?>y=matrix(1:8,ncol=2)>y%*%t(y)


R classes - Matrix

matrixExtracting info

>y[1,] or y[,1]Extending matrices

>cbind(y,seq(101,104))>rbind(y,c(102,109))

Apply is a useful function!>apply(y,2,mean)>apply(y,1,log)


R classes – Data Frame

data.frameCreating it.

> policy.number = c(“A00187”, “A00300”,”A00467”,”A01226”)> issue.age = c(74,30,68,74)> sex=c(“F”, “M”, “M”, “F”)> smoke=c(“S”,”N”,”N”,”N”)> face.amount = c(420, 1560, 960, 1190)> ins.df = data.frame(policy.number, issue.age, sex, smoke,

face.amount)

Accesing it.> ins.df[1,]; ins.df[,1] # access first row, access first colum> ins.df$policy.number # access policy number column> rownames(ins.df); colnames(ins.df);> index.smokers = which(ins.df$smoke == “S”) # row index of

smokers> ins.df[index.smokers] # access all smokers in the df> ins.df$policy.number[index.smokers] # policy number for

smokers


R classes – Data Frame

data.frameManipulating it.

> ins.df = rbind(ins.df, c(“A01495”, 62, “M”, “N”, 1330))> sort.age = sort(ins.df$issue.age, index=T)> ins.df = ins.df[sort.age$ix,]> ins.df$visits = c(0,4,2,1,1)> drops = c(“sex”,”visits”)> ins.df[,!(names(ins.df) %in% drops]> ins.df[,”visits”] = c(0,4,2,1,1)> carins.df = data.frame(policy.number =

c("A01495","A00232","A00187"), car.accident = c("Y","N","N"))> ins.merged.df = merge(ins.df, carins.df, by = "policy.number")> Etc…


R Classes - Factorfactor

Qualitative variables that can be included in models.

>smoke = c(“yes”,”no”,”yes”,”no”)>smoke.factor = as.factor(smoke)>smoke.factor>class(smoke)>class(smoke.factor)


Loops and Conditional Statements

ifExample

>a=9>if(a<0){ print (“Negative number”) } else{ print (“Non-negative number”) }


• for>z=rep(1,10)>for (i in 2:10)

{ z[i]=z[i]+exp(1)*z[i-1] }

• while>n=0>tmp=0>while(tmp<100)

{ tmp=tmp+rbinom(1,10,0.5) n=n+1 }

Loops and Conditional Statements


Functions!• My own functions

> function.name=function(arg1,arg2,…,argN) { Body of the function }

> fun.plot=function(y,z){y=log(y)*z-z^3+z^2plot(z,y)}

> z=seq(-11,10)> y=seq(11,32)> fun.plot(y,z)


Functions! (2)• The ‘…’ argument

– Can be used to pass arguments from one function to another• Without the need to specify arguments in

the header

fun.plot=function(y,z,...) { y=log(y)*z-z^3+z^2 plot(z,y,...) }fun.plot(y,z,type="l",col="red")fun.plot(y,z,type="l”,col=“red”,lwd=4)


Handling data I/O

Reading from files to a data frame>read.csv(“filename.csv“) # reads csv files into

a data.frame>read.table(“filename.txt“) # reads txt files in a

table format to a data.frame

Writing from a data frame to a file>write(x,filename) # writes the object x to

filename>write.table(x,filename) # writes the object x to

filename in a table format

Note: have in mind additional options such as, header = TRUE, row.names = TRUE, col.names = TRU, quotes = TRUE, etc.


Plotting!

>x.data=rnorm(1000)>y.data=x.data^3-10*x.data^2>z.data=-0.5*y.data-90

>plot(x.data,y.data,main="Title of the graph",xlab="x label",ylab="y label")

>points(x.data,z.data,col="red")>legend(-2,2,legend=c("Black points","Red

points"),col=c("black","red"),pch=1,text.col=c("black","red"))


Plotting! (2)

You can export graphs in many formats– To check the formats that are available in your

R installation>capabilities()

png>png("Lab2_plot.png",width=520,height=440)>plot(x.data,y.data,main="Title of the graph",xlab="x

label",ylab="y label")>points(x.data,z.data,col="red")>legend(-2,2,legend=c("Black points","Red

points"),col=c("black","red"),pch=1,text.col=c("black","red"))

>dev.off()eps

> postscript("Lab2_plot.eps",width=500,height=440)


Simulation

Sampling>sample(x,repla

ce=TRUE) – put it back into the bag!

Distributions


Libraries!!

Collection of R functionsthat together perform a specialized analysis.

Install packages from CRAN> install.packages(“PackageName”)

Loading libraries> library(LibraryName)

Getting the documentation of a library> library(help=LibraryName)

Listing all the available packages> library()


• www.bioconductor.org

– A suite of R packages for Bioinformatics.

– To use only Core packages• >source(“http://bioconductor.org/biocLite.R”)• >biocLite()

– To use Core and Other packages• >source(“http://bioconductor.org/biocLite.R”)• >biocLite(c(“pkg1”, “pkg2”,…,“pkgN”))

Libraries!!


Exercise 1 – The empire strikes back: GOOG versus BAIDU

Plot historical Stock Prices times series using prices from yahoo finance.

(a) Download and install tseries package.

(b) Include tseries package as a library in your code.

(c) Use get.hist.quote to download GOOG and BAIDU historical data.

(d) Plot both time series in the same panel and add a legend to the plot.


Exercise 2 – Challenging Challenger

On January 28, 1986, the space Shuttle Challenger exploded in the early stages of its flight. Feynman, along a committee determined that the explosion was due to low temperatures and the failure of O-rings sealed on the booster rockets. The ambient temperature was 36 degrees on the morning of the launch.The scientists had data (temperature, number of failures) from previous flights.


Exercise 2 – Challenging Challenger

(a)Plot the number of failures versus the temperature for flights with one or more O-ring failures. Is there any evidence that temperature affects O-ring performance?

(b)Plot the number of failures versus temperature for all the flights. Is there any evidence that temperature affects O-ring performance?

(c) What’s your conclusion? What do you think the scientists plot before taking the decision to fly that day? Just historical curiosity, Whom played a central role in discovering the causes of the failure and how he announced it?

STAT115 STAT225 BIST512 BIO298 - Intro to Computational Biology Yang Li Lin Liu Jan 29, 2014 1.

Documents

computational biology

computational biology

computational biology4

special files

unzip files wget https

scp username

basic unix command log

files tofrom remote