population mode seems to place results in incorrect order in vcf

[mulderdt]

Hi,

I am running the ont-spectre v0.3.2

I have done the initial single sample calls and get results as expected.

When I run spectre population providing the intermediate files I seem to get the calls listed under the wrong samples. It appears that the sample names are as provided however the entires are in the wrong columns. The columns always seem to provide alphabetical order regardless of input order. If I change the order of the intermediate files to be in alphabetical order I get expected results.

spectre population \
--candidates LR-115_father-F139924.spc.gz LR-115_mother-F139923.spc.gz LR-115_proband-F139921.spc.gz \
--sample-id LR-115 \
--output-dir spectre_output2/ \
--reference hg38_no_alt.fa

gives expected results with expected proband (third column) alteration under proband in third column.

spectre population \
--candidates LR-115_father-F139924.spc.gz LR-115_proband-F139921.spc.gz LR-115_mother-F139923.spc.gz \
--sample-id LR-115 \
--output-dir spectre_output/ \
--reference hg38_no_alt.fa

puts the expected proband (second column) alteration under mother in third column.

spectre population \
--candidates LR-115_proband-F139921.spc.gz LR-115_father-F139924.spc.gz LR-115_mother-F139923.spc.gz \
--sample-id LR-115 \
--output-dir spectre_output3/ \
--reference hg38_no_alt.fa

puts the expected proband (first column) alteration under mother in third column

we can see here the expected call is always in the third position regardless of the order provided and the sample column names.

(base) [dmulder@nextflow01 36b75dfe28818bee712d9dff25e3c3]$ for f in spectre_output*; do echo $f; cat $f/LR-115_spectre_cnv_joint.vcf | grep -v "^##" | grep "^#"; grep 2232000 $f/LR-115_spectre_cnv_joint.vcf; done
spectre_output
#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	LR-115_father-F139924	LR-115_proband-F139921	LR-115_mother-F139923
chr7	2232000	Spectre.DEL.WRJYN5QL	N	<DEL>	.	.	END=2533000;SVLEN=301000;SVTYPE=DEL;CN=1;SUPP_VEC=001	GT:HO:GQ:CN:ID	./.:0.0:0:NULL	./.:0.0:0:NULL	0/1:0.0:60:Spectre.DEL.WRJYN5QL
spectre_output2
#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	LR-115_father-F139924	LR-115_mother-F139923	LR-115_proband-F139921
chr7	2232000	Spectre.DEL.WRJYN5QL	N	<DEL>	.	.	END=2533000;SVLEN=301000;SVTYPE=DEL;CN=1;SUPP_VEC=001	GT:HO:GQ:CN:ID	./.:0.0:0:NULL	./.:0.0:0:NULL	0/1:0.0:60:Spectre.DEL.WRJYN5QL
spectre_output3
#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	LR-115_proband-F139921	LR-115_father-F139924	LR-115_mother-F139923
chr7	2232000	Spectre.DEL.WRJYN5QL	N	<DEL>	.	.	END=2533000;SVLEN=301000;SVTYPE=DEL;CN=1;SUPP_VEC=001	GT:HO:GQ:CN:ID	./.:0.0:0:NULL	./.:0.0:0:NULL	0/1:0.0:60:Spectre.DEL.WRJYN5QL

[oxygen311]

Hi @mulderdt,

Thank you for reaching out.

Currently, ont-spectre does not actively support the population mode and is primarily focused on single-sample germline CNV calling. For use cases involving population-level analyses, we recommend using the original Spectre tool, which may better align with your needs.

If you have further questions or need assistance, please let us know!