dedupe.sh from bbmap excludes contigs with insufficient overlap

0

Entering edit mode

6.8 years ago

biomagician ▴ 410

Hi,

I am using the following command:

grep '>' celegans.fa | wc -l >> numberOfContigsBefore.txt
dedupe.sh -Xmx28g in=celegans.fa out=merged.fa minoverlap=100000000 maxedits=50 minidentity=98 overwrite=t > mergedNotes.txt
    grep '>' merged.fa | wc -l >> numberOfContigs.txt

celegans.fa contains a de novo C. elegans assembly. The genome of C. elegans is 100 MB. I am testing dedupe.sh with an extreme value of overlap to test how it works. My starting number of contigs is 127. Dedupe removes contigs to end up with 109 contigs. I do not understand how this is possible given that I require a minimum overlap of 100 MB and that no contig is that long.

Can anybody explain this to me, please?

Thanks. Best, C.

dedupe bbmap contig assembly • 1.6k views

ADD COMMENT • link 6.8 years ago by biomagician ▴ 410

1

Entering edit mode

You could try to ask to Brian Bushnell, he is the author of BBTools / BBMap.

ADD REPLY • link 6.8 years ago by h.mon 35k

1

Entering edit mode

Looks like the documentation is incorrect; minoverlap is only used when looking for non-containment overlaps (for clustering). I'll either change the behavior or the documentation.

ADD REPLY • link 6.8 years ago by Brian Bushnell 20k

0

Entering edit mode

Hi Brian,

Can you please advise me on how to best use your tool? I would like to remove small contigs that are contained within other contigs and possibly also merge contigs that overlap.

Best,

C.

ADD REPLY • link 6.8 years ago by biomagician ▴ 410

Login before adding your answer.