search

isovic / racon

Ultrafast consensus module for raw de novo genome assembly of long uncorrected reads. http://genome.cshlp.org/content/early/2017/01/18/gr.214270.116 Note: This was the original repository which will no longer be officially maintained. Please use the new official repository here:
https://github.com/lbcb-sci/racon
MIT License
272 stars 49 forks source link

Incorrect indel calling for haploid #171

Closed DewBudd closed 4 years ago

DewBudd commented 4 years ago

I'm using Racon on haploid ONT data and I've noticed that it introduces indels that are not supported by the sequencing data. There will be 12/60 reads that have a deletion at a location and Racon deletes the base but since this is haploid data I would prefer it not to delete that base. I've tried a variety of gap penalties but I've not seen an improvement. Do you have any suggestions?

rvaser commented 4 years ago

Hi Kelly, not sure what causes the indel introduction, you might want to try option --no-trimming which will disable consensus cutting on both window ends. If there are a lot of reads supporting this area, this is the only thing that comes to my mind that could introduce indels.

Best regards, Robert

DewBudd commented 4 years ago

Unfortunately, that did not fix the problem. I'm wondering if it could be a mapping issue. When I view the sam file that I use as input into Racon with IGV, its metric bar at the top will indicate 50% of the reads have a deletion when it is < 25%. Those are the areas where Racon deletes the base.

rvaser commented 4 years ago

Which mapper did you use? Can you share a printscreen of IGV pointing at the problematic area?

DewBudd commented 4 years ago

I've used both minimap2 -ax map-ont and bwa mem -x ont2d with default settings. The screenshot below is from minimap2 but it looks almost identical for the bwa mem alignment. igv_snapshot

DewBudd commented 4 years ago

Apparently IGV was hiding single bp indels and that was the problem. I tried another viewer and the deletion was apparent. Thank you for your help.