No announcement yet.

bowtie 2 dpad parameter explanation

  • Filter
  • Time
  • Show
Clear All
new posts

  • bowtie 2 dpad parameter explanation

    Can someone provide a detailed explanation, what the bowtie2 parameter "--dpad" does? The explanation given in the manual is as follows: "--dpad <int> include <int> extra ref chars on sides of DP table (15)".

    I would interprete it like this:
    1) A sequence string of length <int> is attached to every 5' and 3' end of a read after seeding and global alignment.
    2) Attached sequences are used to solve mapping issues at readends (e.g. deletions present at readends that have been falsley softclipped in previous steps).

    !!!But I am really not confident with this!!!

    I am asking because I realised that increasing dpad to an integer of 100, solves issues I have to face with large deletions in a merged 350bp Illumina amplicon dataset. I mean solving this issue is fine, but I do not have a clue what the possible drawbacks of this setting might be. Furthermore, as I do not exactly understand the parameter and its behaviour, I do not know how to test for possible side-effects.

    Here are my option settings:

    bowtie2 --rdg 5,1 --rfg 5,1 --dpad 100 --score-min L,0,-0.75 --very-sensitive

  • #2
    Sorry the first explanation I posted was from bowties implemented command line help. The following description was actually taken from the manual:
    --dpad <int>
    “Pads” dynamic programming problems by <int> columns on either side to allow gaps. Default: 15.


    • #3
      I sometimes use bowtie2 for the alignment stage of running RSEM (RNA-Seq gene expression). RSEM didn't allow INDELS in the alignments so as part of tweaking bowtie2 to not report those I was advised to set '--dpad 0'.
      /* Shawn Driscoll, Gene Expression Laboratory, Pfaff
      Salk Institute for Biological Studies, La Jolla, CA, USA */