Grep search question

Question

[Omurice]

Hi,

I was wondering if I have 2 sequences of interest: A and B.

Is it possible to do a grep search for reads (in a fastq):

• that contain just sequence "A" and exclude "B" (and vice versa)
• that contain both sequence "A" and "B" together

if so, would someone be able to provide/refer me to the commands to do so?

Thank you so much!

[GenoMax]

Keep in mind that grep will only do perfect matches. You may want to use a NGS data specific tool like seqkit grep or bbduk.sh in filter mode, if you want to allow for other cases than perfect matches.

Answer 1

[Asaf]

You can do it with:

# Grep sequences with A
grep $A file.fastq -B 1 -A 2 > hasA.fastq
# Grep B where A is found
grep $B hasA.fastq -B 1 -A 2 > hasAandB.fastq
# Remove lines in file 1 that are found in file 2
comm -23 hasA.fastq hasAandB.fastq > hasAnotB.fastq

Replace $A and $B with the sequences or set the value of the variables to the sequences you are searching.