Missing calls for all samples except the first in the list

[aguilar-gomez]

Hello,

I ran this command:

jasmine file_list=$VCFlist out_file=$OUTFILE genome_file=$GCA_genome \
    out_dir=$OUTDIR \
    --ignore_strand --mutual_distance \
    --max_dist_linear=0.1 --min_dist=50 --use_end \
    --output_genotypes --normalize_type threads=24

where the VCFLIST contains the path to a vcf with SV variant calls for each individual made with svimasm :

$PATH/AGF7L3/GF_A7_variants_renamed.vcf
$PATH/BGF6L3/GF_B6_variants_renamed.vcf
$PATH/CGF8L3/GF_C8_variants_renamed.vcf
$PATH/DCF1L2/IF_D1_variants_renamed.vcf
$PATH/ECF2L2/IF_E2_variants_renamed.vcf

Each of these files looks like this:

bcftools view $PATH/BGF6L3/GF_B6_variants_renamed.vcf |grep -v "##"|more|cut -f1-3,6- |head
#CHROM	POS	ID	QUAL	FILTER	INFO	FORMAT	GF_B6
CM063496.1	111037	svim_asm.INS.1	.	PASS	SVTYPE=INS;END=111037;SVLEN=68	GT	1/0
CM063496.1	122761	svim_asm.INS.2	.	PASS	SVTYPE=INS;END=122761;SVLEN=4857	GT	1/0
CM063496.1	152390	svim_asm.INS.3	.	PASS	SVTYPE=INS;END=152390;SVLEN=612	GT	1/0
CM063496.1	155910	svim_asm.DEL.1	.	PASS	SVTYPE=DEL;END=155978;SVLEN=-68	GT	1/0
CM063496.1	214293	svim_asm.DEL.2	.	PASS	SVTYPE=DEL;END=214596;SVLEN=-303	GT	1/0
CM063496.1	254990	svim_asm.DEL.3	.	PASS	SVTYPE=DEL;END=255032;SVLEN=-42	GT	1/0
CM063496.1	268856	svim_asm.INS.4	.	PASS	SVTYPE=INS;END=268856;SVLEN=442	GT	0/1
CM063496.1	290646	svim_asm.DEL.4	.	PASS	SVTYPE=DEL;END=296947;SVLEN=-6301	GT	0/1
CM063496.1	312271	svim_asm.INS.5	.	PASS	SVTYPE=INS;END=312271;SVLEN=71	GT	0/1

The output I got from jasmine only has the variants from the first sample in the list (GF_A7):

bcftools view jasmineOUTPUT.vcf | grep -v "##" | cut -f1-3,9-| head

[W::vcf_parse_info] INFO 'STRANDS' is not defined in the header, assuming Type=String
[W::vcf_parse_info] INFO 'CHR2' is not defined in the header, assuming Type=String
#CHROM  POS     ID      FORMAT  0_GF_A7 1_GF_B6 2_GF_C8 3_IF_D1 4_IF_E2
CM063496.1      53276   0_svim_asm.DEL.1        GT:IS:OT:DV:DR  0/1:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      415840  0_svim_asm.INS.8        GT:IS:OT:DV:DR  0/1:.:INS:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      418364  0_svim_asm.DEL.8        GT:IS:OT:DV:DR  0/1:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      418461  0_svim_asm.DEL.9        GT:IS:OT:DV:DR  0/1:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      419822  0_svim_asm.INS.9        GT:IS:OT:DV:DR  0/1:.:INS:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      479781  0_svim_asm.DEL.10       GT:IS:OT:DV:DR  0/1:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      562838  0_svim_asm.DEL.12       GT:IS:OT:DV:DR  1/0:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      771827  0_svim_asm.DEL.20       GT:IS:OT:DV:DR  0/1:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
CM063496.1      788179  0_svim_asm.DEL.22       GT:IS:OT:DV:DR  0/1:.:DEL:0:.   ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA ./.:NA:NA:NA:NA
[W::vcf_parse_info] INFO/END=30185055 is smaller than POS at CM063496.1:39985045

All samples have ./.:NA:NA:NA:NA as genotype except GF_A7, even when in the input files all the samples have valid genotypes.

What am I doing wrong?

Thank you,
Diana

[sjteresi]

Hello,

I also came across this issue this morning while working on my samples, searched the GitHub issues and came across your description. I think I may have determined what is going on, so I figured I'd chime in.

I installed with conda in mid-January 2026, so our software versions should be similar.
I used only slightly different settings from you, but I imagine the pipeline outputs are still comparable:

INPUT_LIST="$1"
OUTPUT_VCF="$2"
THREADS=${3:-2}

export _JAVA_OPTIONS="-Xmx100g"

/usr/bin/time -v conda run -n jasmine \
jasmine \
    file_list=${INPUT_LIST} \
    out_file=${OUTPUT_VCF} \
    max_dist=500 \
    --nonlinear_dist \
    --allow_intrasample \
    --output_genotypes \
    threads=${THREADS}
bgzip -f -@ ${THREADS} ${OUTPUT_VCF}

At least in my data it appears as though the variants are not strictly sorted by position in the output (there appears to be some sort of pattern of writing blocks of data that are sorted by both position and the reformatted variant ID), once I sorted the VCF it became clear that I did indeed have variants genotyped in Sample Number 2 and beyond. Though that still doesn't explain why there are all of those missing genotype calls.

Good luck,
Scott