如何使用Linux命令在Fasta文件中提取标头的一部分

Question

I have a Fasta file with unique header,I would like to extract a part of this header by using Regular expression in Unix.

for example My Fasta file start with this header:

>jgi|Penbr2|47586|fgenesh1_pm.1_#_25  

and I would like to extract just the last part of this header like:

>fgenesh1_pm.1_#_25

Actually I use this regular expression in vim editor but It did not work:

:%s/^([^|]+\|){3}//g

or

:%s/^([A-Z][0-9]+\|){3}//g

I would be appropriate if give me some suggestion.

Answer 1

You can use sed :

sed -e 's/>.*|/>/' fasta-file

ie everything between > and | is replaced by > .

Answer 2

I don't know if the leading > is also a part of your text. Assume that they are not.

Since you tagged with vim , then I just post the vim solution.

You can make the usage of the "greedy" of regex:

In vim:

%s/.*|//

will leave the last part, this is the easiest way.

in vim you can use \\zs, \\ze and non-greedy matching too:

%s/\zs.\{-}\ze[^|]\+$//

Of course, if you like grouping, you can use \\(...\\) to group and don't use \\zs \\ze .

In your codes, you grouped just with (...) without escaping. I don't know how did you configure your magic setting in your vimrc, if you use default, you have to escape the ( and ) to give them special meanings (grouping here). Just like what we do with BRE. Do a :h magic , and find the table to know the difference.

In vim do :h terms to get detail information.