Deblur Paired End Reads

Hello there,

I have paired-end sequences (16S, V4). I would like to denoise with 2 ways:

  • joined paired-end reads
  • forward and reverse reads (separate)

I have 3 questions:

  1. It seems that Deblur only supports forward reads at this moment. Is it best to combine forward and reverse reads for Deblur, or just Deblur forward reads?
  2. To denoise separately for forward and reverse reads, will it work if I Deblur the reverse reads similarly to the forward reads?
  3. Also, in general, what are the implications of (possibly) neglecting the reverse reads?

I hope my questions make sense and I’m not being too messy with the terminology. Thank you in advance for the clarification!


I think the answer to that depends on your goals, and how “joinable” the reads are. I would suggest attempting with joined reads first, and if that doesn’t work for you, try to work with only the forward reads.

I’m not 100% sure, but I suspect in theory this would work. @wasade?

Shorter reads can make it harder to resolve taxonomy.

1 Like

Thanks, @thermokarst!

Deblur is agnostic to whether the data are joined or not. If you decide to join them it would need to be performed upstream of Deblur.

Deblur’ing forward and reverse reads separately has not been explored to the best of my knowledge. If it works, it would probably result in a large loss of sequence data when stitching as your ability to pair is contingent on the pairs existing on both outputs.

Longer reads aren’t assured to improve taxonomic resolution and may increase the total error present in the data.



This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.