0

I just recently started to use powershell in win7 in order to produce pipeline like scripts for the program mothur. Before I used bash scripting in ubuntu to do this. I am happy that everything works well now except one task:

I would like to like to format a fasta file that is in the form:

filename.fasta:

>HXXC990
AGTTCAAGGTCTCT
>HXXC991
GGGTTTCAAATCTC
>HXXC992
GGGTCTCTCCTATA

To a file that is tab-delimited and looks like that

output.file:

HXXC990    filename
HXXC991    filename
HXXC992    filename

It is important that the first column of the output file contains the names without the ">"-signs. and the second by tab delimited column the original filename.fasta without the suffix ("filename"). I have the solutions gci to read out the base name of the file and Select-String to output all the lines beginning with ">". The only problem remains the formatting in the two columns and the constant repetition of the file name in the second column.

I've tried so far:

Select-String '>' .\filename.fasta | % {$_.Line} | set-content output.txt

to produce a file containing only the lines that contain the ">" signs. Afterwards I just replaced them. The file name I've got by

$base1 = gci filename.fasta | % {$_.BaseName}
www
  • 38,575
  • 12
  • 48
  • 84
user2251977
  • 15
  • 2
  • So I've tried "Select-String '>' .\filename.fasta | % {$_.Line} | set-content output.txt" to produce a file containing only the lines that contain the ">" signs. Afterwards I just replaced them. The file name I've got by $base1 = gci filename.fasta | % {$_.BaseName}. – user2251977 Apr 06 '13 at 11:25
  • I've taken the liberty to update your question with your code snippets. – Ansgar Wiechers Apr 06 '13 at 12:38
  • Sure. Thanks. That was my fault, sorry! Thanks again! Future posts will be now definitely in the right format. – user2251977 Apr 06 '13 at 15:47

2 Answers2

0

Try this:

select-string '^>' filename.fasta | % {
  $_ -replace '^.*\\(.*?)\.fasta:\d+:>(.*)$', "`$2`t`$1"
} > output.file

Note that your regular expression should be ^>, not just >. The latter would match > anywhere in a line.

This can be applied to more than a single file like this:

$recurse = $false

Get-ChildItem "C:\base\folder" -Filter *.fasta -Recurse:$recurse `
  | select-string '^>' `
  | % { $_ -replace '^.*\\(.*?)\.fasta:\d+:>(.*)$', "`$2`t`$1" } > output.file
Ansgar Wiechers
  • 193,178
  • 25
  • 254
  • 328
  • Thanks a lot! That was exactly I was looking for. I still have to get used to the syntax. I used to use awk for these kind of operations. But in a windows environment, it is more elegant to use the powershell. Thanks a lot again! – user2251977 Apr 06 '13 at 15:44
0

Here's another solutions, showing some different options for the operations involved:

gci *.fasta | select-string '^>(.+)' |
 % {"{0}`t{1}" -f $_.matches.groups[1],$_.filename.split('.')[0]} |
 Set-Content output.file
mjolinor
  • 66,130
  • 7
  • 114
  • 135