Extract genes names from GOrilla output file

Question

[rrapopor]

Hello, I got a file with GOrilla results. I don;t have the input data used to create the file. This is the file:

GO Term Description     P-value FDR q-value     Enrichment      N       B       n       b       Genes
GO:0001580      detection of chemical stimulus involved in sensory perception of bitter taste   4.40E-20        6.67E-16   10.14   20504   48      1053    25      [Tas2r105  -  taste receptor, type 2, member 105, Tas2r119  -  taste receptor, type 2, member 119, Tas2r136  -  taste receptor, type 2, member 136, Tas2r122  -  taste receptor, type 2, member 122, Tas2r117  -  taste receptor, type 2, member 117, Tas2r123  -  taste receptor, type 2, member 123, Tas2r115  -  taste receptor, type 2, member 115, Tas2r129  -  taste receptor, type 2, member 129, Tas2r130  -  taste receptor, type 2, member 130, Tas2r125  -  taste receptor, type 2, member 125, Tas2r121  -  taste receptor, type 2, member 121, Tas2r124  -  taste receptor, type 2, member 124, Tas2r120  -  taste receptor, type 2, member 120, Tas2r113  -  taste receptor, type 2, member 113, Tas2r114  -  taste receptor, type 2, member 114, Tas2r109  -  taste receptor, type 2, member 109, Tas2r110  -  taste receptor, type 2, member 110, Tas2r106  -  taste receptor, type 2, member 106, Tas2r107  -  taste receptor, type 2, member 107, Tas2r102  -  taste receptor, type 2, member 102, Tas2r116  -  taste receptor, type 2, member 116, Tas2r104  -  taste receptor, type 2, member 104, Tas2r103  -  taste receptor, type 2, member 103, Tas2r131  -  taste receptor, type 2, member 131, Tas2r140  -  taste receptor, type 2, member 140]

I want to get a list with genes only by extract from the brackets in the 11th column the genes name. For example:

Tas2r105
Tas2r119
Tas2r117
...

I tried the code :

awk -F'[][]' '{print $2}' gorilla_master_nox.txt | grep -oP '(?<=,).*?(?=-)'

But I dont get the wanted results. I would appreciate any help.

Thank you

[cpad0112]

If gene nomenclature pattern is fixed, you can use:

$ grep -Po '\w{3}\d\w\d{3}' test.txt 

Tas2r105
Tas2r119
Tas2r136
Tas2r122
Tas2r117
...

If you want to tighten the expression, you can use [A-Z][a-z]{2}[0-9][a-z][0-9]{3}. However, if you are not sure of gene pattern, please use cut -f11 test.txt| grep -Po '[A-Z][a-z]{2}[0-9][a-z][0-9]{3}'

[rrapopor]

Thanks! The genes nomenclature pattern is not fixed, but the cut -f11 didn't work.

Answer 1

[rpolicastro]

Here are a grep (with Perl) and pure Perl solution If you want to get all genes within the brackets for each line. I sort the results and only keep one of the unique values using sort -u so feel free to remove that part if you don't want that behavior.

grep with Perl enabled.

grep -oP '\w+(?=\s{2}-)' gorilla_master_nox.txt | sort -u

A (minutely slower) Perl one liner.

perl -nle 'print "$1" while /(\w+)\s{2}-/g' gorilla_master_nox.txt | sort -u

The first 5 genes of the results using either method.

Tas2r102
Tas2r103
Tas2r104
Tas2r105
Tas2r106

[cpad0112]

I am aware that first part of grep in OP tightens the expression. However, double space,hyphen,double space might be unique in file. Given that, grep expression can be modified.

$ grep -oP '\w+(?=\s{2}-\s{2})' test.txt 
Tas2r105
Tas2r119
Tas2r136
Tas2r122
Tas2r117
Tas2r123
...

If it is not unique, we can use cut -f11 test.txt| grep -oP '\w+(?=\s{2}-\s{2})

[rpolicastro]

Thanks, to reduce it even more double space-hyphen-double space gave me the same results as double space-hyphen on the example data, so I used the later.

[rrapopor]

Thank you! For the entire file I got different results for the grep and the perl solutions. The perl solution worked perfect for me:)