modkit pileup --phased return only combined haplotype while processing whole genome

[JbaptisteLam]

Hi modkit team,

modkit version: 0.6.4

cmd:
modkit pileup
--phased
--bgzf
--header
--threads 24
--modified-bases '5mC'
--ref hg38.fa
barcode1.bam
barcode1.modkit.phased

I'm trying to generate phased bedmethyl files using '--phased' option on haplotagged bam with longphase.

Note that, BAM comes from adaptive sampling experiment on a few sample targeting only chrX

After testing different args combinations I saw that without setting '--region', pileup created only the combined.bed.gz no matter other args like --cpg or --combine-strands, while hp1 and hp2 output remain empty.
I would expected no result on autosomes and some data on chrX for both haplotype

With --region 'chrX' everything goes well, both haplotype and combined file were filled.

NB: HP tag is correclty set, I double checked with samtools and on IGV

Am I missing something ?

Best,

JB

[JbaptisteLam]

It seems that this behaviour only happend when reference is not set

[ArtRand]

Hello @JbaptisteLam,

Could you expand on what you mean by "reference is not set"? You have --ref in your command above and --modified-bases will require --ref. Could you paste the commands that worked and the ones that didn't?