[SOLVED] manipulating text in linux enviroment

Question

[theodore]

Dear all.

I do not know if a tool already exists but I would like to do the following steps in a tab delimited file:

The file is as follows:

chr1 0 600 15 Repetitive/CNV 0 . 0 600 245,245,245
chr1 1000 1600 8 Insulator 0 . 1000 1600 10,190,254
chr1 100004000 100005200 2 Weak Promoter 0 . 100004000 100005200 255,105,105
chr1 100005200 100016800 13 Heterochrom/lo 0 . 100005200 100016800 245,245,245
chr1 10001600 10014800 13 Heterochrom/lo 0 . 10001600 10014800 245,245,245
chr1 100016800 100022800 12 Repressed 0 . 100016800 100022800 127,127,127
chr1 100022800 100026800 13 Heterochrom/lo 0 . 100022800 100026800 245,245,245
chr1 100026800 100028600 12 Repressed 0 . 100026800 100028600 127,127,127
chr1 100028600 100037000 13 Heterochrom/lo 0 . 100028600 100037000 245,245,245
chr1 100037000 100046600 12 Repressed 0 . 100037000 100046600 127,127,127
chr1 100046600 100046800 6 Weak Enhancer 0 . 100046600 100046800 255,252,4
chr1 100046800 100047000 2 Weak Promoter 0 . 100046800 100047000 255,105,105
chr1 100047000 100047200 4 Strong Enhancer 0 . 100047000 100047200 250,202,0
chr1 100047200 100047400 6 Weak Enhancer 0 . 100047200 100047400 255,252,4
chr1 100047400 100054200 13 Heterochrom/lo 0 . 100047400 100054200 245,245,245
chr1 100054200 100055000 12 Repressed 0 . 100054200 100055000 127,127,127
chr1 100055000 100087400 13 Heterochrom/lo 0 . 100055000 100087400 245,245,245
chr1 100087400 100087600 6 Weak Enhancer 0 . 100087400 100087600 255,252,4

first I would like to remove the number space before the characterization of the area: 6 Weak Enchancer ---> Weak Enchancer second to count all Weak enchancer or other identical fields of row 4 and print something like the following: Weak Enchancer 4 Heterochrom 20 . . . I tried: sort 'file.bed' | awk '{print $4}' | uniq -c -D -i or sort 'file.bed' | uniq -c -D -i with no avail. Any help will be higly appreciated

I should state that I want to do it as easily as possible, I have no real skills in programming and even if openoffice can do it I'm fine with that!!!

Thank you in advance

Theodore

Answer 1

[Devon Ryan]

cut -f 4 file.bed | sed 's/[0-9]* //' | sort | uniq -c

or something along those lines.

Edit: For the sake of clarity, the cut -f 4 portion extracts column 4 and the sed command just replaces a number followed by a space with nothing (i.e., it removes that pattern).

[theodore]

I've run the pipeline, it works great, although I get the following:

45 10_Txn_Elongation
    133 11_Weak_Txn
     95 12_Repressed

it seems as if sed had replaced spaces (\s) with underscore (_)???

[Devon Ryan]

It shouldn't do that, at least not unless you changed it to be something like sed 's/ /_/'.

Answer 2

[chefer]

Using your original file, you can do the unique count (col 5 in the example) like this:

cut -f 5 test.bed | sort | uniq -c

Which gives you this:

6 Heterochrom/lo
1 Insulator
1 Repetitive/CNV
4 Repressed
1 Strong Enhancer
3 Weak Enhancer
2 Weak Promoter

edit: I misunderstood the example format, this will not work on the input dataset. You should cut on column 4, and then do the replace as suggested by the upvoted post.

[Devon Ryan]

It's understandably unclear from the question, but 2 Weak Promoter appears to be an example of a single column value, rather than being split into two columns. It appears you split things into different columns when you made the file on your local system.

[theodore]

It is a tab delimited file the 4th row consists of (number)(space)(description). The description is with space values.

I do not know how to make a tab look like in a copy paste manner.

[Neilfws]

Some ways to paste or represent tab in the terminal discussed here.

[theodore]

I'll try it tomorrow. Thank you

Answer 3

[theodore]

cut -f 9 | sed 's/[0-9]* //' | sort | uniq -c | sed 's/* //' | awk '{print $2,$3"\t"$1}'

I got it thanks to your recommendations.

The above command/pipeline worked miracles for me