Duplicated reads (IDs) from nanopore sequencing
1
0
Entering edit mode
14 days ago
njornet ▴ 20

Hello,

I basecall the pod5 files generated by the sequencer with dorado and align with minimap2 and in the BAM file I get the exact same read with the same ID and all more than once. Is this normal? If it is, I suppose I have to remove the duplicates right?

Thank you!

bam nanopore sequencing • 435 views
ADD COMMENT
0
Entering edit mode

I get the exact same read with the same ID and all more than once

If you have short(er) reads you could be seeing secondary alignments. Which would be a normal thing. Check the SAM tags

I don't think I have seen what @marco.barr is referring to below.

ADD REPLY
0
Entering edit mode

I had the same problem in this iusse https://github.com/nanoporetech/dorado/issues/603 and I solved it as they say here

ADD REPLY
0
Entering edit mode

Thanks for the link. We never use "pod5_fail" folders when re-basecalling so we did not see this issue. Current version of MinKNOW now makes a single pod5 folder.

Since original question is about alignments following minimap2 showing duplicate ID, it may likely be due to secondary alignments.

ADD REPLY
0
Entering edit mode

it's also true that now I only have one pod5 folder, thanks anyway for clarifying the focus of the issue. Yes, they are probably secondary alignments

ADD REPLY
0
Entering edit mode

I think this is the answer but I can't find in the tags info about primary and secondary alignments. I've only found in the flag that the second entry for each repeated ID corresponds to a supplementary alignment, but from what I've seen that's not the same a what you are saying.

ADD REPLY
0
Entering edit mode
ADD REPLY
0
Entering edit mode

If you only have supplementary alignments then they may be caused by reasons mentioned here --> What are Supplementary Reads? I did not mention them originally since they are not main reason for dup id's.

If the alignments are valid you can't simply ignore them as being from duplicate ID's.

ADD REPLY
0
Entering edit mode

I've looked into it and one of the reads with a supplementary alignment has the representative alignment aligning to the + strand and being 1207 bp long and the supplementary aligning to the - strand being 130 bp, but they are overlapping so this is not the indel thing I think

Supplementary Alignments
chr4:433,255-434,368 (-) = 1,115bp @MAPQ 60 NM132 (representative)
chr4:434,228-434,367 (+) = 140bp @MAPQ 33 NM20 (supplemantary)

Here is a screenshot of both reads in IGV, the representative in red and supplemetary in blue.

enter image description here

I've found this thread but they are talking about bwa and I don't think they got anywhere.

ADD REPLY
0
Entering edit mode
14 days ago
marco.barr ▴ 140

Hi, I'm also working with ONT data and with dorado. From experience I tell you that some versions of MInKNOW have bad management of fast5 files, wrote duplicate skip folders as both pod5 and fast5. Just filter on unique reads id:s and it should be fine. However I also solved it using updated versions of dorado and MInKnow. They fixed this problem with new versions of dorado and MInKnow. What versions are you using?

ADD COMMENT

Login before adding your answer.

Traffic: 2782 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6