LinuxQuestions.org - [SOLVED] replace multiple patterns from a file

- Linux - Newbie (https://www.linuxquestions.org/questions/linux-newbie-8/)

- - replace multiple patterns from a file (https://www.linuxquestions.org/questions/linux-newbie-8/replace-multiple-patterns-from-a-file-949117/)

replace multiple patterns from a file

Hi all,

I have a fasta file that looks like this

Code:



>consensus_5353 

 ACTCGGAGGTCCGATCCAAAGTTTTTCTTTTCAGTGCCGAGTAGAGTTAC

>consensus_5354 

 GCCGGTGGTGGCGGAAGCTGGTGGTTGGCCGGCCGGGTCGAGTGGAGGTC

>consensus_5355 

 AGCAGGAGTCCCGCCGCCCTCGACCTCTCTTCTTCCGCCGCCGCCCGGTG

>consensus_5356 

 AATTAGTGGATTATTTGGTGAGGTGGATGTAGAGTGTAGACCGTATAATT

>consensus_5357 

 GCAATACTCAAATTGGAATAGGATGAGCAAGGAAGAGGAAAATGGTGGGG

>consensus_5358 

 TAAGATGTTCTGTTAGGGACTCAGAAGAACATCAGAATCACTACTTACGT

>consensus_5359 

 TGCTGGACTTGCTCTTAGCCTCCGATCGTCCCTATAGACTTTTGGCCTTT

>consensus_5360 

 ACGTGCCGCACGTGGGGTACAAAACCACCGCGGCGTAGGAGACGTCAAAA

>consensus_5361 

 TCGGCGGGGAGCAGCTAGTAACGCGCATTAACACGAGCAAATCCTAGAGA

I have another file with the list of patterns. My pattern list file looks like this

Code:

>consensus_5353

>consensus_5357

>consensus_5359 

>consensus_5361

What I want to do is to look for the patterns in my fasta file and delete those patterns and the next line of sequence corresponding to that pattern. Any thoughts how do I go about it with sed or awk or grep or perl?

I knew how to use sed to delete a single pattern but how to delete multiple patterns from a file instead of writing something like this

sed -e '/pattern/d' -e '/pattern/d' file

Thanks in advance.

sed s/pattern//g input > output

Probably not the best way to do it but it works.

Here is a script that might help.

Please note that I have used sed with "-i" so it will change your actual file. It is preferable you try this script on a backup file to make sure things don't go wrong.

Code:

 

#!/bin/bash

while read line

do 

sed -i "/$line/ N;s/$line.*//" fasta

done < pattern

Quote:

Originally Posted by bsat (Post 4698464)

Code:

 

#!/bin/bash

while read line

do 

sed -i "/$line/ N;s/$line.*//" fasta

done < pattern

Many thanks for the replies. The above script solved my query.