csq errors with phase problem

[jd3234]

Hi,

I am suddenly experiencing the same problem as described in this issue a couple years ago: https://github.com/samtools/bcftools/issues/1628

Strange thing is that it used to work with the gff3 before and even with the same bcftools version (1.21) but I rebuild it in between (using conda...) the last successful run and now. I suspected some conda problems and built the latest version from scratch (as described here https://samtools.github.io/bcftools/howtos/install.html) but problem still occurs.

Using /bcftools/misc/gff2gff does not help.

Any idea what might be causing the issue?

[pd3]

Can you share the GFF, or part of it, and show the command you are using please?

[jd3234]

Hi, thanks for looking into this.

This is my command and the error message:

bcftools csq -f ChineseLong-9930_v3.fasta -g ChineseLong-9930_v3.gff3 --local-csq --ncsq 32 minimal.vcf
Parsing ChineseLong-9930_v3.gff3 ...
Warning: Ignoring GFF feature with unknown phase .. chr1	FIXBIOTYPE	exon	22415	22939	.	-	.	ID=CsaV3_1G000010.1.exon3;Parent=CsaV3_1G000010.1
Note: truncated transcript CsaV3_1G000280.1 with incomplete CDS (this is very common)
Error: GFF3 assumption failed for transcript CsaV3_4G021870.1, CDS=12946538: phase!=len%3 (phase=2, len=-1). Use the --force option to proceed anyway (at your own risk).

and this is the gene model that causes trouble

chr4	.	gene	12943582	12947020	.	+	.	ID=CsaV3_4G021870;Name=CsaV3_4G021870
chr4	.	mRNA	12943582	12947020	.	+	.	ID=CsaV3_4G021870.1;Parent=CsaV3_4G021870;Name=CsaV3_4G021870.1;biotype=protein_coding
chr4	.	exon	12943984	12943984	.	+	2	ID=CsaV3_4G021870.1.exon1;Parent=CsaV3_4G021870.1
chr4	.	CDS	12943984	12943984	.	+	2	ID=CsaV3_4G021870.1.cds1;Parent=CsaV3_4G021870.1
chr4	.	exon	12946538	12946706	.	+	1	ID=CsaV3_4G021870.1.exon2;Parent=CsaV3_4G021870.1
chr4	.	CDS	12946538	12946706	.	+	1	ID=CsaV3_4G021870.1.cds2;Parent=CsaV3_4G021870.1

Like I said the really strange thing is that this exact gff3 used to work just a couple month ago (last time in April, around August I rebuild the container and since then it its not working anymore). Unfortunately I have no installation of bcftools from before that time.

I can confirm that the GGF3has not been changed since 2023 as it is integrated into a system where no one except of the admin has any rights to change it.

[pd3]

Can you make available the entire GFF file, not just the snippet?

[jd3234]

Hi, I am attaching the full GFF3. Thanks!

ChineseLong-9930_v3.gff3.txt

[jd3234]

I would also share the fasta but it is too big to put it here. Let me know if you need it then we can discuss another way to share it.