This is a test version of Biostars. For the public version, visit https://www.biostars.org.
PCR duplicata and de novo assembly ??

Hi,

Sorry I am a beginner in NGS field I used SGA preqc to assess the quality of my data.

This program estimates the PCR duplicate proportion to 20%.

1) First, I am not sure to understant what is PCR duplicate, what is the difference with the sequencing coverage ?

2) 20%: Is it a lot ?

3) If yes, how to remove them ? is it important prior to a de novo assembly ?

pcr duplicates preqc
• 3,194 views • link
updated by Medhat Medhat • written by Picasa Picasa

2 answers

There is two types of duplication PCR duplication and optical duplication, we remove duplicates mainly to reduce recurrent errors.

  • PCR duplication are introduced during library preparation

    you can find nice info about it here also

  • optical duplicates (Illumina) are obtained when a single cluster of reads is part of two adjacent tiles' on the same slide and used to compute two read calls separately

here you can use this tool to remove them
Note: while using the tool I suggested

To remove the duplicate records from the resulting file, set the REMOVE_DUPLICATES parameter to true

• 0 views • link
written by Medhat Medhat

Thanks for your support.

1) Did you have any good experience with removing those duplicates ? (your N50 improved a lot )

2) Should I remove it on raw reads or trimmed reads ?

• 0 views • link
written by Picasa Picasa
  • from the word duplication leaving them will not change N50 "cause actually they are the same no information will be introduced by leaving them" on the other hand it will introduce bias in downstream analysis.
  • first you trimme your reads then remove them
• 0 views • link
written by Medhat Medhat

20% duplicates are not a lot. And I think for assembly they should not matter much. These are duplicate reads in your sample. For quantitative analyses such as ChIP-seq and RNA-Seq, they can pose a big problem.

• 0 views • link
written by Satyajeet Khare Satyajeet Khare

Log in to answer this question.

More posts like this