pwd # Tell me which directory I'm in
ls # What else is in this directory
ls .. # What is in the directory above me
ls foo/bar/ # What is inside the bar directory which is inside the foo/ directory
ls -lah foo/ # Give the the details (-l) of all files and folders (-a) using human
# readable file sizes (-h)
cd ../.. # Move up two directories
cd ../foo/bar # Move up one directory and down into the foo/bar/ subdirectories
cp -r foo/ baz/ # Copy the foo/ directory into the baz/ directory
mv baz/foo .. # Move the foo directory into the parent directory
rm -r ../foo # remove the directory called foo/ from the parent directory
find foo/ -name "*.gff" # find all the files with a gff extension in the directory foo/less bar.bed # scroll through bar.bed
grep chrom bar.bed | less -S # Only look at lines in bar.bed which have 'chrom' and
# don't wrap lines (-S)
head -20 bar.bed # show me the first 20 lines of bar.bed
tail -20 bar.bed # show me the last 20 lines
cat bar.bed # show me all of the lines (bad for big files)
wc -l bar.bed # how many lines are there
sort -k 2 -n bar.bed # sort by the second column in numerical order
awk '{print $1}' bar.bed | sort | uniq # show the unique entries in the first columngrep foo bar.bed # show me the lines in bar.bed with 'foo' in them
grep foo baz/* # show me all examples of foo in the files immediately within baz/
grep -r foo baz/ # show me all examples of foo in baz/ and every subdirectory within it
grep '^foo' bar.bed # show me all of the lines begining with foo
grep 'foo$' bar.bed # show me all of the lines ending in foo
grep -i '^[acgt]$' bar.bed # show me all of the lines which only have the characters
# a,c,g and t (ignoring their case)
grep -v foo bar.bed # don't show me any files with foo in themawk '{print $1}' bar.bed # just the first column
awk '$4 ~ /^foo/' bar.bed # just rows where the 4th column starts with foo
awk '$4 == "foo" {print $1}' bar.bed # the first column of rows where the 4th column is foo
awk -F"\t" '{print $NF}' bar.bed # ignore spaces and print the last column
awk -F"\t" '{print $(NF-1)}' bar.bed # print the penultimate column
awk '{sum+=$2} END {print sum}' bar.bed # print the sum of the second column
awk '/^foo/ {sum+=$2; count+=1} END {print sum/count}' bar.bed # print the average of the
# second value of lines starting
# with foogrep -hv '^#' bar/*.gff | awk -F"\t" '{print $1}' | sort -u
# grep => -h: don't print file names
# -v: don't give me matching files
# '^#': get rid of the header rows
# 'bar/*.gff': only look in the gff files in bar/
# awk => print the first column
# sort => -u: give me unique values
awk 'NR%10 == 0' bar.bed | head -20
# awk => NR: is the row number
# NR%10: is the modulo (remander) of dividing my 10
# awk is therefore giving you every 10th line
# head => only show the first 20
awk '{l=($3-$2+1)}; (l<300 && $2>200000 && $3<250000)' exercises.bed
# Gives:
# contig-2 201156 201359 gene-67 24.7 -
# contig-4 245705 245932 gene-163 24.8 +
# Finds all of the lines with features less than 300 bases long which start
# after base 200,000 and end before base 250,000
# Note that this appears to have the action before the pattern. This is
# because we need to calculate the length of each feature before we use it
# for filtering. If they were the other way around, you'd get the line
# immediatly after the one you want:
awk '(l<300 && $2>200000 && $3<250000) {l=($3-$2+1); print $0}' exercises.bed
# Gives:
# contig-2 201156 201359 gene-67 24.7 -
# contig-2 242625 243449 gene-68 46.5 +#!/usr/bin/env bash
set -e # stop running the script if there are errors
set -u # stop running the script if it uses an unknown variable
set -x # print every line before you run it (useful for debugging but annoying)
if [ $# -ne 2 ]
then
echo "You must provide two files"
exit 1 # exit the programme (and number > 0 reports that this is a failure)
fi
file_one=$1
file_two=$2
if [ ! -f $file_one ]
then
echo "The first file couldn't be found"
exit 2
fi
if [ ! -f $file_two ]
then
echo "The second file couldn't be found"
exit 2
fi
# Get the lines which aren't headers,
# take the first column and return the unique values
number_of_contigs_in_one=$(awk '$1 !~ /^#/ {print $1}' $file_one | sort -u | wc -l)
number_of_contigs_in_two=$(awk '/^[^#]/ {print $1}' $file_two | sort -u | wc -l)
if [ $number_of_contigs_in_one -gt $number_of_contigs_in_two ]
then
echo "The first file had more unique contigs than the second"
exit
elif [ $number_of_contigs_in_one -lt $number_of_contigs_in_two ]
then
echo "The second file had more unique contigs"
exit
else
echo "The two files had the same number of contigs"
exit
filess or head to double check their format-F"\t"history and it might have a recordman the_name_of_a_command often gives you helpIf you wanted me to calculate the sum of all of the scores for genes on contig-1 in a bed file, I'd probably run each of the following commands before moving onto the next:
head -20 bar.bed # check which column is which and if there are any headers
head -20 bar.bed | awk '{print $5}' # have a look at the scores
awk '{print $1}' bar.bed | sort -u | less # check the contigs don't look wierd
awk '{print $4}' bar.bed | sort -u | less # check the genes don't look wierd
awk '$4 ~ /gene-/' bar.bed | head -20 # check that I can spot genes
awk '($1 == "contig-1" && $4 ~ /gene-/)' bar.bed | head -20 # check I can find
# genes on contig-1
# check my algorithm works on a subset of the data
head -20 bar.bed | awk '($1 == "contig-1" && $4 ~ /gene-/) {sum+=$5}; END {print sum}'
# apply the algorithm to all of the data
awk '($1 == "contig-1" && $4 ~ /gene-/) {sum+=$5}; END {print sum}' bar.bedYou should probably use awk if:
You should probable use grep if:
grep -r foo bar/: look in all the files in bar/ for any with the word 'foo')You should use find if:
You should write a script if:
You should probably use less or head: