U NIVERSITY OF THE W ITWATERSRAND , J OHANNESBURG

Wits Bioinformatics

Introduction to Linux

Scott Hazelhurst
University of the Witwatersrand
www.bioinf.wits.ac.za/courses/linux

February 2015

Contents
1 Introduction 1

2 Linux – first look 2

3 Command line interface 3

3.1 Interacting with Un*x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4 Files and directories 7

4.1 A word on file and directory names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.2 Manipulating directories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.3 Directory/file information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.4 Manipulating files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.5 Copying and renaming files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.6 Manipulating directories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.7 Further manipulation of files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.8 Permissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5 Process Control 15

6 Additional and advanced material 17

6.1 Environmental variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

7 Pipes and Redirection 19

7.1 Combining processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
7.2 Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
7.3 xargs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

1 Introduction
Material can be found at www.bioinf.wits.ac.za/courses/linux

1
Operating system function
Acts as a software layer to provide access to the underlying hardware
• Higher-level of abstraction

• Sharing of resources

• Memory

• Files

• Processing power

• Communication

• Protection

Examples
Unix-like
• Unix, AIX, Solaris

• Free BSD

• GNU/Linux – Ubuntu, Chrome, . . .

• MacOS X
Windows
• XP

• Windows 7,8
Lots of others

Historical
• IBM 370; OS/2
Mobile
• OSE Symbian
Safety-critical systems:
• Integrity

• OSEK

2 Linux – first look

In these notes, I will refer to Linux and Unix interchangeably. Of course, this is a gross simplification that
will upset lots of people. Most of what is covered here holds for operating systems that are inspired by
the original Unix system of AT&T. Apple’s Mac OSX and Linux are probably the best known examples.
Sometimes Linux is called GNU/Linux (also controversial) because although the underlying operating
system is Linux, usability relies on many tools that were developed as part of the GNU project.
2
Why GNU/Linux?

• Good quality

• Free

• Many software packages run on it.

• Most common for bioinformatics (and many other scientific areas)

Distributions
Many distributions (flavours) of Linux

• Different look and feel

• Some of the systems programs and environments different

• Mostly for a user interchangeable

Examples
Ubuntu, RedHat, Fedora, Suse, Debian, Scientific Linux, Centos, Mandrake
Can download from www.mirror.ac.za

Running multiple operating system

• Dual boot

• Multi-boot

• Virtualisation

– VirtualBox
– Parallels/VMWare

3 Command line interface

Command line interface/shell

3
Many different CLI shells – allows interation with OS.
• bash/sh
• csh/tcsh
• many others
On the whole very similar Usually a program like Terminal or xterm that provides access
• Many shell languages – can be used as a programming language

GUI
• Easier to learn
• More intuitive
• Easier mental models
• Memory cues
• Quicker for many things
CLI
• you may not have an option
• More powerful, control
• Great for repetitive tasks
4
Example 1: run water

CLI Equivalent

./water -asequence prots.fa -bsequence sampl.fa \

-gapopen 10 -gapextend 5 \
-outfile data.aln

Example 2
Move file Documents/january.txt to Data/feb.txt

• GUI: Click, drag, drop

• mv Documents/january.txt Data/feb.txt

Example 3
Suppose you have directories /opt/data/exp/YY/text/local/control

• In each directory, there are files xxxx-YY-month-ddd.xxxx

Copy the files xxxx-YY-mar-ddd.eXXX from all these directories into a directory /tmp/exp_data/
march

• GUI?

• CLI:

cp /opt/data/exp/*/text/local/control/*-mar*.e* \
/tmp/exp_data/march

5
Example 4
You have new data in a file myseq.fa and a directory db containing 1875 files.

• Run the water program 1875 times to compare your myseq.fa file against each of the files in db in
turn.

3.1 Interacting with Un*x

On modern Unix systems such as GNU/Linux there is reasonably friendly graphical interface. On your
screen you are likely to have a windows open. Example windows mights be: a web browser, an editor,
some applications, and most importantly a shell.
The shell is a window in which you can enter commands to the Unix system – the shell is the interface
between you and the system. You may have a number of shells running at the same time.
It is easy to use the shell: just enter the command you want and (a) the system does what you want,
or (b) it doesn’t and complains that it doesn’t understand you.
The conventional syntax of commands issued by the user to a shell is:

command command_options command_parameters

The command options and ordering of command parameters may differ slightly from version to version
of Unix so it is best to use the on-line help provided by Unix to determine the exact available options and
parameter order.
A small note on syntax convention. In Unix, the fullstop character does not have any significant
purpose within file names and may appear several times in a single file name (and need not occur at all).
However,

• To make life easier for ourselves we adopt a conventions of naming files – do .py is used for Python
programs, .c is used for C programs, .pdf is used for PDF files, .tex is used for LATEX files, and so
on. Binary executables often do not have a suffix.

• Many program use these conventions to guess the contents of files. So a program like Firefox may
use the suffix to guess which external program to use to display a file. The LATEX program expects
its main input files to have .tex suffixes. But these are conventions that are not enforced by the
operating system. Using sensible conventions is as much for your benefit than for the computer’s.

• File names beginning with a fullstop have significance to Unix because they contain environment
and configuration information useful to both the system and user. Normally files that are start with
a “.” are not shown when you list the directory and some times not in the GUI file browser.

Command options are also preceded by minus signs to distinguish them from command parameters. For
example, wc data.txt says count the number of characters, words and lines in the file data.txt. But if we
include the option -l then we only count the number of lines: wc -l data.txt If a file name is required
as a parameter to a command and is not provided, the shell will by default use the standard input and
output i.e. the terminal.
We’ll look at what command you can enter and how you interact with the shell in this section.

3.2 Getting started

Initial interaction with the system depends on whether it’s the computer in front of you, whether it’s on
the other side of the world, whether there’s a GUI, whether there is security.

6
Getting started
Normally you log-in

• (Normally), enter user id password to log in

• If running a GUI, run a terminal program

Once terminal runs

• Enter commands

• Case-sensitive

• Kill terminal: Control-D.

Need to log out of session.

Changing password
The standard Un*x command for changing passwords is passwd command though there are variants for
networked systems.

• Choose sensible passwords

Command history
To see a list of the most recent commands issued to the shell:

• history

To repeat a previous command from this list, type:

• !number

where number is the number of the command.

Command completion
If you press the tab key, shell tries to complete as much of the command or file name as possible.

• If you type fi then TAB, all commands starting with fi shown.

• If you type ls /usr/sh followed by TAB, system can complete to /usr/share

• If you type ls /usr/l followed by TAB, system is not able to complete further as there are several
options. But if you press TAB twice all the options shown.

Command editing

• The left and right cursor keys, and the delete key allow you to edit the current command;

• Control-A moves to the start of the current command;

• Control-E moves to the end of the current command; and

• The down and up cursor keys allow you to move forward and backward in history.

7
On-line help
• apropos
e.g., apropos music
• man
e.g., man passwd
• info
e.g., info ls

4 Files and directories

All information in Unix is stored in files. The Unix file system is organized so as to allow multiple users to
maintain their files in a single hierarchy.
This hierarchy can be viewed as a single tree structure of directories and files with links that may
connect to directories and files anywhere else in the structure.
File system
Hierarchical collection of directories and files
Each file has:
1. Contents
2. A Name
3. A location (Directory path)
4. Administrative info. (Who owns it, how big it is, date — see example below)
• The collection of files/directories in a Linux system can be thought of as an inverted tree.
• Top of file system tree is the “Root”, represented by /
• Can have cross-tree links

In the GNU/Linux file system tree each user has a unique location to work called their HOME directory.
Users are automatically placed in their home directory when they login.

4.1 A word on file and directory names

• A space is a delimiter. Avoid spaces in names.
• forward slash / — separates dir/file names
• backward slash \ treat the next character differently to usual.

lpr cats and dogs

lpr cats\ and\ dogs

• Avoid: $, ~ and quotes in names

• Files starting with . are hidden
• Don’t start a file name with a –
It’s free country: you can flout all these conventions. But don’t.
8
 var/
etc/

httpd/

Play/
your-home-directory/
ROOT = /

students/
home/

Work/
usr/
Figure 1: Linux directory Tree

4.2 Manipulating directories

Current working directory
CWD is place in file hierarchy where session is.

• When you first log on you will be in your HOME directory;

• Use the cd to change current directory

cd dirpath

• pwd command shows CWD.

Many commands by default assume you mean the current working directory.

The system administrator can set up their system so that home directories are where convenient for
the organisation

• On Linux system’s, often user bob is in /home/bob

• On MacOS X, /Users/bob

• But often variations – safest is to refer to ~bob or the environment variable $HOME

E.g., the ls command lists the files in the directory:

• ls : lists the file in the current working directory;

• ls -l : (small L) lists files in the current directory with a number of details for each file;

• ls /usr/share : lists the contents of directory /usr/share.

There are many other options for ls – you can use the man page to find out.

9
Paths
Each file has a path – where the file is in the file system.

• Path can be absolute as in /usr/share/emacs/23.4/etc/JOKES

File name is JOKE & path is /usr/share/emacs/23.4/etc

Path can be relative to the current working directory

• JOKES
No path given – current working directory is implied

• funny/JOKES
The file JOKES in the directory funny that is in the current directory

• funny/very/JOKES
and so on. . .

Special paths

• ~
Tilde: home directory of current user

• ~scott
Home directory of user scott

• ..
Parent directory (relative)

• .
Current directory (relative)

4.3 Directory/file information

Directory/file info
Following information kept

• The name

• The name of the owner of the file

• The name of the group of the file

• The date the file was last changed

• The permissions of the file.

• other stuff

The ls -l command shows this

10
4.4 Manipulating files
manipulating files

Examining files

1. cat

2. more

3. nano, emacs, vim

4. head -n 25 fname

5. tail

You can use cat filename to see the contents of an entire text file on the standard output (terminal).
On the other hand, more filename allows you to view the entire text file on the standard output one
screen at a time (at the more prompt, a space character will scroll a full screen down while a carriage
return character will scroll one line down).

4.5 Copying and renaming files

Copying files
To create a copy of a file, use the command:

• cp source dest

• cp a.txt b.txt

• cp a.txt /data/dir

• cp -r data backup

The destination can either be a file (with a path name) or a directory. If it a directory a copy of the file is
made and put in the destination giving the copy the original name. If the destination is a file then a copy
is made of the source and the copy is given the name of the destination.
Note that by default cp does not copy directories. You must use the -r option (or another recursive
option). Like many commands, cp has many options. Doing a info coreutils ’cp invocation’ will
show you these.

Moving a file/Renaming a file:

To move a file from one location to another, use the command:

mv source dest

If the filename in the second parameter is different, a rename is achieved.

Deleting file
rm

• rm data.dat
11
4.6 Manipulating directories
Creating, deleting subdirectories
mkdir

• mkdir newdir

• mkdir /usr/local/other

• mkdir newdir/subdir

• mkdir -p newdir/subdir/subsub/other

To delete a directory you use the rmdir command.

• must be empty

To delete a directory and all its contents, you must first delete the files in the directory using rm and then
use rmdir. An alternative approach is to use rm -r for recursive delete. This a very powerful and very
dangerous option. There is hardly a system administrator in the world

Changing directory
To change the current working directory, do: cd newdirpath

• Change to the directory /usr/share/: cd /usr/share/

• Change to user’s home directory: cd ~

• Change to parent of cwd: cd ..

• Change to the directorty site-lisp in cwd

cd site-lisp

• cd ../gnumeric

cwd == current working directory.

4.7 Further manipulation of files

The head and tail commands allow you to see the top and bottom of the file.
• head snps.txt show the first 10 lines of the file;

• head -n20 snps.txt show the first 20 lines of the file;

• head -n-3 snps.txt show everything except the last 3 lines (this doesn’t work on standard MacOS
X);

• tail snps.txt show the last 10 lines of the file;

• tail -n30 snps.txt show the last 30 lines of the file;

• tail -n+3 snps.txt show everything except the first lines;

• tail -f snps.txt show the last 10 lines of the file and then wait for further input. This is useful if
you have a program that is writing to a file, and in another shell you want to monitor the output.
12
Extracting colums and rows
Being able to extract out interesting from a file is important.
• cut: extract columns from a file. There are two basic modes (you can read about others in the man
file). Columns can be extracted based upon horizontal position (numbering columns by character,
using the -b option. Or, columns can be extracted by field where the different columns are assumed
to separated by field delimiter (by default a tab).

cut – extracting columns

cut -b 10-20,30 logs.txt
cut -f 1,2,5,7 results.csv
cut -f 1,2 -d, results.csv
cut -f 1,2 -d\ results.csv

The examples above allow us to extact and manipulate files with data. We can do the same thing very sim-
ply using a program called Excel and with small files this is probably easier because the GUI allows more
intuitive interaction. But consider a realistic example where there are 1000 rows and 10000 columns.
Such a large file would be very slow and clumsy to manipulate through Excel or similar program.
The grep command can be used to extract rows based upon what they contain.
grep – extracting lines that match

grep rs637812 data.map

grep rs637812 data.map results.map
grep -ir bioinformatics Teaching

• Show context
grep -C 2 rs837812 *map

• Count number of occurrences

grep -c NA317813 *fam

• Which file
grep -H NA317813 *fam

• grep -f patterns.txt *
Use the lines in the files patterns.txt as the things to grep for.

Instead of just using plain text files, grep also allows you to search for regular expressions, but this is an
advanced topic we are not covering now.
There are other powerful tools that can be used too – awk and sed are very powerful tool that allows
extracting and manipulating files.
There are many useful ways of combining files, including paste and join.

4.8 Permissions
File access is determined by the file’s protection status. You may control access to your files and directories
by granting and denying access privileges to either the user (you), the group the user belongs to (the pg
group for example) or all other users. These privileges are read, write and execute.
13
Permissions
Privileges specified for

• user (owner)

• group

• other

Privileges are:

• read

• write

• execute (files) enter (directory)

Part of a listing of files in a directory may look like this (using ls -l):

-rw-r--r-- 1 jayesh pg 110 Nov 16 13:11 theory.tex

drwx------ 8 jayesh pg 512 Mar 12 1992 Personal
-rwxrx-r-- 1 jayesh pg 230 Aug 27 12:52 a.out

Here we have three files: theory.tex, Personal and a.out. In each case the owner is the user jayesh and the
group is pg.
The 10 characters on the left indicate file protection and is interpreted in the following manner. An r
indicates that the user can read the file and w means the user can change the file. The x stands for execute.
For ordinary files, an x indicates that this file stores a program that is ready to run and can be run. For
directories, x indicates permission to enter the directory.

1. character 1 – if there is a dash the file is an ordinary file; if it is a d, this is a directory; if it is a l this
is a link.

2. Characters 2–4 indicate the permissions of the owner of the file.

3. Characters 5–7 indicate the permissions of members of the group.

4. Characters 8–10 indicate the permissions of other users.

In the example above, the user jayesh can read and change the file Theory.essay, members of the pg group
and indeed all other users can just read it. The file Personal is actually a directory. jayesh can read and
write to this directory, and also enter the directory. Noone else can access the directory. The file a.out can
be read, written and executed by jayesh. Members of the pg group can read and execute the file, while all
other users can just read it.

Changing permissions
chmod changes permissions on any files.

• chmod g+rwx file1 will give members of the group associated with the file the ability to read, write
and execute it.

• Or chmod o-rx file1 will remove read and execute privileges or other users.

• chmod ug+x file1 gives the owner (user) and group permission to execute.
14
 • chmod a+rwx will give all users permissions to read, write and execute files.

• chmod o=r file gives others the ability to read the file and removes any other permissions that others
may have.

• chmod g=rx,o=r file gives group ability to read and execute, others the ability to read the file and
removes any other permissions previously had

Directories:
• For directories, x means that the permission holder can enter the directory.

• The permission X means: if the file is a directory, then x, otherwise ignore.

chmod a+X *

Numeric permission:
• 4: read permission. 2: write permission. 1: execute permission.

• 5=read and execute. 3=write and execute. 7=all

• Typically given as as triple RGO:

752

Exercise: Give your own home directory the following permission.

• user: read, write, execute

• group: read, execute

• other: execute
For each directory in your home directory set the permissions so that you can read, write and execute
but that no-one else has any permissions. The only exception is public_html which you should give read
and execute privileges to all users.
There are more complex permissions possible too in the standards permission model together with
access control lists. But this is beyond our course.

5 Process Control
Process control
Many jobs can execute at the same time (same, different users)

• foreground

• background

Jobs may belong to currrent shell, or others

• Each job as a PID (global)

• Each job has job number (specific to that shell) Use % to refer to this
15
Viewing jobs
Jobs in current shell

• jobs shows job number

• ps shows PID

See other jobs

• ps -u scott

• ps -a

Manipulating jobs of current shell

If the job is in the foregrond

• Control-C kills

• Control-Z suspends

Can reactivate suspended jobs

• fg put it foreground

• bg put in background

• bg %3 (job 3 – if several suspended jobs)

Killing any job

kill pid
kill 3617
kill -9 3617

Running a job
To run a job in the foreground give the command:

emacs
python hello.py
xcal
xeyes

To start in the background put an ampersand afterwards

xeyes &
emacs newprog.py &

16
Monitoring Load

• top
The top command shows the processes that are using most of the CPU. There’s a whole lot of other
info you can get too.

• The w command shows you who’s logged on, as well as current load.

top also shows you memory usage. This can often be critical for system performance. The most
important columns in this respect are %MEM and RES.

Exercise: How do you monitor the load given by a specific process.

Aside – discovering resources

• Load needs to be compared to the resources on the system

• /proc/cpuinfo

• /proc/meminfo

screen – window manager for command-line

screen allows separation of terminal from physical device

• multiple virtual terminals on same device;

• move terminal session from one physical device to another;

• robust against time-outs

Particulary useful for remote access

screen has many features. In this section, I am only showing a selection of the key features, but some very
useful features are being omitted.
Creating a new screen

screen -S name

Creates a new terminal session, putting existing session in background

• terminal session gets a unique number

• named by some user-defined meaningful identifier.

• terminal session destroyed when you do a C-d

• terminal session detached with C-a d

17
Listing terminal sessions
The command is screen -ls
Note that each terminal has the name you gave it and a unique number.
Reattaching terminal sessions
You can reattach a session to any terminal.
screen -r name
One slightly annoying feature of screen is that it uses C-a as the default escape key sequence (e.g., C-a d
detaches the current session). This is annoying because C-a is a commonly used key sequence in editing
commands (go to the beginning of the current line). You can chang the behaviour by using the -e option.
For example,
screen -e^Mm -s update
will create a new terminal session called update, and instead of C-a being the escape sequence, C-m willb.

6 Additional and advanced material

Globbing
Ways of specifying many things at the same time
• *
• []
Wildcard
• cat *
• ls a*.dat
• ls *mar*.e*
• /bin/rm hello-201[01].[oe]xx

Gaining super-powers
Actions allowed determined by permissions
• Some actions only allowed for root user
• Some users can be authorised to act for root

sudo
ls /root

sudo ls /root

Becoming someone else

su bob

Must know bob’s password (or be root)

18
Getting super-powers for a session

# if you know root’s password

su

# if you are a sudoer -- use your password

sudo su

It is strongly recommended to use sudo rather than su. This limits inadvertent damage you can do to your
system.

6.1 Environmental variables

Each shell has a set of environment variables

• Used by programs to guide behaviour

• Environment vars set per-terminal – modifying one doesn’t modify others

• Init set up in /etc/bash.bashrc (system-wide) and .bashrc (personal)

• To see environment variable: printenv or printenv VARNAME

• To use an environment variable – prefix with $, e.g. echo $HOME

• Set environment variables using:

export PATH=/opt/local/bin:${PATH}
(note no dollar on left-hand side)

The PATH environmental variable tells the system where to look for executables. These directories are
searched in order, and the first path found with the named executable will be used.
Typically the current working directory is not on the PATH. That is why when you want to execute a
script in the current working directory that you have to say ./myscript rather than just myscript This
helps (a little) in preventing inadvertent execution of scripts.

Key environment variables

• PATH: binaries

• PYTHONPATH

• R LIBS

• PERL5LIB

• PYTHONPATH

• HOME

• HOSTNAME / HOST

• USER
19
Initialisation
Environment variables (and other initialisation) can be set by start up scripts

• /etc/profile

• .bash profile : login shells

• .bashrc : other interactive shells

Can use the source command to load another initialisation script.

7 Pipes and Redirection

7.1 Combining processes
Most of the commands we have seen so far produce output on the terminal; some may also take input
from the terminal. Unix allows the terminal to be replaced by a file for either or both of input and output.
This becomes useful for interaction required by background processes as well as for allowing users to
combine commands to create their own commands.
Input-output redirection and pipes are provided by Unix for this purpose. Input-output redirection
The syntax for output and input redirection are given by the following forms respectively:

Redirection

command >filename
command < filename

The symbol > means “put the output in the following file, rather than to the terminal” and the symbol
< means “get the input to this command from the following file, rather than from the terminal”.
>> used instead of > appends output to the file rather than overwriting contents of the file.

For example, the command sequence

ls > /tmp/wdirfiles
wc -l < /tmp/wdirfiles

can be used to count the number of files in the working directory.

There are more advanced features too.

• >> can be used to append to a file.

• A common use of cat and redirection is to combine file.

cat *.dat > all.dat

cat dates*.csv | sort | uniq > all_dates.txt
20
7.2 Pipes
Pipes
Pipes allow you to glue together programs

• output of one program becomes input of the next

ls | wc -l

It is common practice to put the output of one program into the input of another via a temporary file as in
the example above. This can be achieved by first using a command with an output redirection followed by
one with an input redirection. However, by doing so we incur a storage overhead cost for a temporary file
to store the input to the second program. Furthermore, it would be more efficient to run both programs
in parallel so that a continual output from the first program can be fed into the second program. This
observation leads to one of the fundamental contributions of the Unix system, namely the pipe.
A pipe (denoted by a vertical bar i.e.. |) is a way to connect the output of one program to the input
of another without any temporary file; a pipeline is a connection of two or more programs through pipes.
All programs in the pipeline execute in parallel to achieve good performance. Only data dependencies
restrict the flow of data between these programs. Any program that reads from the terminal can read
from a pipe instead and any program that writes on the terminal can write to a pipe. This is where the
convention of reading the standard input when no files are named pays off: any program that adheres to
the convention can be used in pipelines. grep and sort are two examples often used in pipelines.
For example, the command: who | grep mary | wc -l counts the number of times user mary is
logged on.

Example
List in alphabetic order the capitals of the countries which use the West Africa CFA franc.
Get the data file and inspect it
Now extract column data and sort

grep "West African CFA" results.csv |

cut -f 2 | sort

7.3 xargs
Converts standard input to arguments.
Suppose we have a file that contains status of various files

january.dat GOOD
feb.dat BAD
march.dat BAD
april.dat GOOD
may.dat BAD
...

How can we delete the bad files?

Find the bad files
21
 • grep BAD statusfiles.csv
Extract out the names
• grep BAD statusfiles.csv | cut -f 2
Delete them
Extract out the names
• grep BAD statusfiles.csv | cut -f 2 | xargs /bin/rm

Programming in bash
Can type commands in shell, or save in file and run
• e.g., doexps.sh
• hash-bang
• make executable
• run like so ./doexps.sh
Normally, an executable program is expected to be machine code that can run directly on the machine. If
the program is a script in bash or some other language, you need to tell the system. One way of doing it
is to explicitly tell it:
• bash doexps.sh
But it is useful both to save typing and more importantly to help users of your scripts to have a way of
telling the system how to interpret your file. Recall that the name of the file does not have this information.
The hash-bang is a special sequence on the first line that specifies how the script should be executed. For
example, a bash script would have
• #! /bin/bash
and if it were a Python program
• #! /usr/bin/python
Example script
#! /bin/bash
N=10
BASE="gwas14"
BED=${BASE}.bed
BIM=${BASE}.bim
FAM=${BASE}.fam
plink --bfile sample --bmerge $BED $BIM $fam --make-bed --out xxx

Can pass a parameter

Example script
#! /bin/bash
N=10
BASE=$0
BED=${BASE}.bed
BIM=${BASE}.bim
FAM=${BASE}.fam
plink --bfile sample --bmerge $BED $BIM $fam --make-bed --out xxx
22