How To Extract And Quantify Duplicated Rna-Seq Reads?

Question

[Pals]

I am working on a RNA-Seq data. The Fastqc shows that data has duplication level of >65%. Can anyone tell me how I can extract the duplicated reads and quantify them? I am also interested to rank those duplicated reads.

Thanks!

[Irsan]

Why do you want to extract and quantify them? They are already quantified by FastQC right? Are you worried about the big number? Then you should Duplicated Reads In Rna-Seq Experiment where it is explained that in RNA-seq many of the reads come from ribosomal RNA and highly expressed genes resulting in a high duplication level in your reads

[Pals]

just curious to see the actual reads that are duplicated thousands of times.

[Irsan]

gunzip -dc yourFastq.gz | awk '{if(NR%4==2)print $0}' | sort | uniq -c

[Pals]

Thank you very much Irsan. I had to slightly modify your trick awk '{if((NR-2)%4==0)print $0}' :-)

[Irsan]

Cool, I changed the comment

[Pals]

Unfortunately, sorting did not go well. I want the most repetitive sequence ranked, where this command sorts alphabetically.

[Zev.Kronenberg]

print the reads and then | sort | uniq -c

[Pals]

Thank again, finally its done. gunzip -dc yourFastq.gz | awk '{if(NR%4==2)print $1}' | sort | uniq -c | sort -g

[Zev.Kronenberg]

glad you go it working.

Answer 1

[Zev.Kronenberg]

After marking duplicates (with your favorite program) you can use the samtools flag to pull out the reads you want.

Just set the include flag: samtools view -f 0x400:

0x400 PCR or optical duplicate