Modified bam results

[partha434]

Dorado produces basecalls in unmapped bam file. How can I check the modifications from this bam file?

[sklages]

Map your data on your reference genome using minimap2^[1] and then run modbam2bed^[2] on the alignment file. You can check the results in e.g. IGV^[3]

^[1] https://github.com/lh3/minimap2
^[2] https://github.com/epi2me-labs/modbam2bed
^[3] https://software.broadinstitute.org/software/igv/

[partha434]

But this is an unmapped bam file. How can I map a bam file to reference genome?

[Psy-Fer]

You can go ubam to fastq and carry the meth tags with something like this
samtools fastq -TMM,ML ${USAM} | minimap2 -x map-ont -a -t ${num_threads} -y --secondary=no ${REFERENCE} - > ${SAMFILE}

The magic here is the - T that tells it to carry the MM and ML tags to the fastq, then the -y in Minimap2 tells it to carry them back to the bam.

Then you can use modbam2bed and other downstream tools.

[cjw85]

If you do:

samtools fastq -T '*' ....

you can carry all the tags from the SAM without loss.

[iiSeymour]

@partha434 we are working on producing aligned BAMs for the next release.

Note: you need at least samtools 1.16 (Aug 2022) for samtools fastq -T '*' https://github.com/samtools/samtools/releases

[partha434]

Thank you so much for the codes. This really worked for me... @Psy-Fer @cjw85
Eagerly waiting for an update related to this on Dorado @iiSeymour

[iiSeymour]

v0.3.0 adds the ability to output aligned BAM directly from the basecaller as well as dorado aligner which can map/remap uSAM/uBAM and retain tags.

https://github.com/nanoporetech/dorado#alignment