tool for a pattern of repeats

Question

[2nelly]

Dear all,

I am wondering if there is any tool or someone have a custom script to extract reads from sam/bam or fastq file that contain a specific pattern of repeats. Let s say we have 1000000 reads and we want to look how many of them contain the pattern (CCAATT)n.

Thank you in advance.

Answer 1

[Pierre Lindenbaum]

using samjs: http://lindenb.github.io/jvarkit/SamJavascript.html

java -jar dist/samjs.jar -e 'record.getReadString().contains("CCAATT") || record.getReadString().contains("GGTTAA")'  input.bam

using awk:

samtools view -h input.bam | awk -F '\t'  '($0 ~ /^@/ || $10 ~ /CCAATT/ || $10 ~ /AATTGG/ )'

from a fastq:

  gunzip -c input.fastq.gz | paste - - - - | awk  -F '\t'  '($2 ~ /CCAATT/ || $2 ~ /AATTGG/ )' | tr "\t" "\n"

[2nelly]

Thank you so much Pierre!

You are always there when we need help.

I will give a try to all of these commands

[Pierre Lindenbaum]

You've asked many questions on this forum without validation (green mark on the left) or comment. e.g: samtools flagstat only for exome coordinates . Please validate/close the answers.

[2nelly]

Dear Pierre,

These commands are working nicely.

However, what about if we want to look for reads consisting of let s say more than 50% of this repeat (randomly distributed within the reads)?

many thanks