meta data for this page
  •  

This is an old revision of the document!

Basic bash commands

We provide below a set of basic commands to be used in the shell together with an example application. To test the effect of these command, download the file ejemplo.txt below, and give it a try. Note, the example file seems to contain useless information. Still, once you understood what the commands are doing, they will help you during the course to extract and format information from sequence files, even if they contain millions of lines.

ejemplo.txt
XXXXXXXXXXXXX
aaaaa	xxxxx
xxxxx	bbbbb
ccccc	xxxxx
xxxxx	ddddd
eeeee	xxxxx
aaaaa	bbbbb
ddddd	fffff
axaxa	bxbxb
XXXXXXXXXXXXX

Play with the following commands. If necessary, make changes to find out what they are doing. Remember to call the file only in the first command given: cat ejemplo.txt | grep “xxxxx”

Reading files + counting lines

  • cat ejemplo.txt
  • less ejemplo.txt (press q to exit)
  • head -n3 ejemplo.txt (returns the first three lines of the file)
  • tail -n2 ejemplo.txt (returns the last two lines of the file)
  • cat ejemplo.txt | wc -l (counts the number of lines in the file)
  • cat ejemplo.txt | uniq (displays the content of the file by removing successive identical lines)
  • cat ejemplo.txt | sort | uniq (sorts the content of the file, removes successive identical lines, and displays the output)
  • cat ejemplo.txt | sort | uniq -c (counts the number of unique lines in the file)

Pattern matching using grep

  • grep “xxxxx” ejemplo.txt (search for the pattern 'xxxxx' in the file ejemplo.txt)
  • grep “xxxxx” ejemplo.txt | wc -l (counts the number of lines in ejemplo.txt that contain the pattern 'xxxxx')
  • grep -c “xxxxx” ejemplo.txt (also counts the number of lines in ejemplo.txt that contain the pattern 'xxxxx')
  • grep -v “x” ejemplo.txt (returns the lines that do not contain the pattern)
  • grep -i “x” ejemplo.txt (makes the pattern matching case insensitive)
  • grep -o “aaaaa” ejemplo.txt (returns only the matching pattern)
  • grep -A2 “c” ejemplo.txt (returns the matching plus the next two lines)
  • grep -B2 “c” ejemplo.txt (returns the matching plus the two preceding lines)
  • grep -E “aaaaa|ccccc” ejemplo.txt (returns lines that contain either 'aaaaa' or 'ccccc'

Exercise

Exercise

What would you have write to find lines that contain both “a” and “b”?

Regular expressions

  • grep “^x” ejemplo.txt (returns all lines starting with an 'x')
  • grep “x$” ejemplo.txt (returns all lines ending with an 'x')
  • grep “aaaaa…” ejemplo.txt (How many characters are returned?)
  • grep “a*” ejemplo.txt (returns all lines)
  • grep “a.*” ejemplo.txt
  • grep “a[^a]” ejemplo.txt (returns all lines containing a pattern where at least one 'a' is followed by a different character)
  • grep “a[ax]” ejemplo.txt (returns all lines where an 'a' is either followed by an 'a' or by an 'x')

Note that the regular expressions work on “file contents”. Do not confuse with wildcards.

Exercise

Exercise

Regular expressions can be combined with the previous options, particularly with -o to extract characters after a pattern. Try extracting only the first 3 characters of lines that start with “a”

Text editing using sed

sed s/x/i/ ejemplo.txt
sed s/x/i/g ejemplo.txt
sed s/axaxa/kkkkk/ ejemplo.txt

Sed and special characters. (An special character might be a space, a tab(\t), a symbol reserved for regular expressions ($), etc.): 

sed s/\t// ejemplo.txt
sed -e 's/\t//' ejemplo.txt

Exercise

Exercise

Combine with regular expressions → Try converting the first 3 characters into “iii”

Cut

Extract columns from a table 

cut -f2 ejemplo.txt

Exercise

Exercise

Extract the first column of lines that contain the character “d”

Save/overwrite output

(ACHTUNG: do not give same name as input file) 

cut -f2 ejemplo.txt > copy_ejemplo.txt
To save without overwriting (lines get added to file):
grep "X" ejemplo.txt >> copy_ejemplo.txt

Other commands

Echo: Print something on the terminal 

echo "Witness me"

Translate: Convert one character to another. 

cut -f1 ejemplo.txt
cut -f1 ejemplo.txt | tr -d '\n'
Notice anything in the output? Allow me to fix it:
cut -f1 ejemplo.txt | tr -d '\n' | sed -e 's/$/\n/'

Rev: print each line backwards 

echo "a b c d e"
echo "a b c d e" | rev

Common errors

Calling the file twice

Calling the file twice 

grep "XXX" ejemplo.txt | sed -e 's/X/V/g' ejemplo.txt

The output of the first part of the command is not getting passed as the input to the second part of the command, since the file is being read again.

Moving files

Moving files 

cat ejemplo.txt > copy.txt
grep "aaa" copy.txt > copy.txt
cat copy.txt

Input file is also the output. The contents of input file will be brutally deleted.

Grep and quotations

Grep and quotations 

echo -e $">Important_sequence\nAAATTCTCACCCCTCAGAAA" > important_sequence.fasta
cat important_sequence.fasta
grep > important_sequence.fasta

The crocodile “>” without quotations will be interpreted as “save file” instead of “find”. The original file will be overwritten with nothing.

Wildcards or regular expressions

Wildcards or regular expressions 

sed -e 's/a*//' ejemplo.txt

Note that “*” has a different meaning as a wildcard and as a regular expression. Wildcard “*” = Regex “.*”

File names

File names

Some naming schemes can make your life as a programmer a living hell.

output file final final definitive 6.fasta
aaa
bbb
ccc
rename -e 's/_/ /g' output_file_final_final_definitive_6.fasta
cat output\ file\ final\ final\ definitive\ 6.fasta

Part II

For-loop

First we will create another file. Observe differences in output when concatenating both files and when working with them separately in a for-loop: 

grep "fffff" ejemplo.txt > copy_ejemplo.txt
cat ejemplo.txt | wc -l
cat copy_ejemplo.txt | wc -l

Both files 

cat *plo.txt | wc -l

For-loop

  • for i in *plo.txt; do echo $i; done
  • A list of files that you want to process; namely, the files ejemplo.txt and copy_ejemplo.txt
  • Instruction for each element of your list; in this case, print said element on the terminal. “$i” is each element of list “i”; note that the variable name is specified by the user, so I could create the list “k” and instruct each element “$k”.
  • The rest is part of the basic sintax of a for-loop.

We next change the instruction from the example above, to reading the file and counting its lines: 

for i in *plo.txt; do cat $i | wc -l; done

Which file is which? 

for i in *plo.txt; do cat $i | wc -l | sed -e "s/^/$i\t/"; done