Issue with VCF (does not represent a pangenome graph)

Hi @eblerjana,

I am running pangenie, using the HPRC-CHM13 (88 haplotypes) dataset VCF files, as you've specified here.

However, I am getting an error message suggesting that the VCF is not suitable:

Determine allele sequences ...
Read reference genome ...
Found 25 chromosome(s) from the reference file.
Read input VCF ...
terminate called after throwing an instance of 'std::runtime_error'
  what():  GraphBuilder: variant at chr1:11714 overlaps previous one. VCF does not represent a pangenome graph.

I have copied my runscript below for context. I am running version 3.0.1 of pangenie

#!/bin/bash -l

#SBATCH --job-name=pangenie
#SBATCH --account=pawsey0933
#SBATCH --partition=work
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=8
#SBATCH --mem=16G
#SBATCH --nodes=1
#SBATCH --time=00:30:00
#SBATCH --mail-user=chiara.folland@perkins.org.au
#SBATCH --mail-type=END
#SBATCH --error=%j.%x.err
#SBATCH --output=%j.%x.out
#SBATCH --export=ALL

# Load modules
module load singularity/4.1.0-slurm
module load bcftools/1.15--haf5b3da_0
module load samtools/1.15--h3843a85_0 #htslib

# Set variables 
input_file=/scratch/pawsey0933/cfolland/pangenie/samplesheet_test.csv
image_dir=/scratch/pawsey0933/cfolland/pangenie/container/
image_name=/scratch/pawsey0933/cfolland/pangenie/container/pangenie.sif
ref=/scratch/pawsey0933/cfolland/pangenie/refs/CHM13v11Y.fa
graph_vcf=/scratch/pawsey0933/cfolland/pangenie/refs/chm13_cactus_filtered_ids_biallelic.vcf
outdir=/scratch/pawsey0933/cfolland/pangenie/output
decomp_script=/scratch/pawsey0933/cfolland/pangenie/scripts_gs/convert-to-biallelic.py

# Read the input file and process each sample
awk -F, 'NR>1 {print $1, $2, $3}' $input_file | while read -r sample fq1 fq2; do
    echo "Processing sample: $sample"

    # Run PanGenie
    singularity exec -B /scratch/pawsey0933/cfolland/pangenie ${image_name} PanGenie \
        -i <(zcat ${fq1} ${fq2}) \
        -r ${ref} \
        -v ${graph_vcf} \
        -t 23 \
        -j 23 \
        -o ${outdir}/${sample}

    # Decompose bubbles
    cat ${outdir}/${sample}_genotyping.vcf | python3 ${decomp_script} ${graph_vcf} \
        > ${outdir}/${sample}_genotyping_biallelic.vcf

    # Collect vcf stats
    bcftools stats ${outdir}/${sample}_genotyping_biallelic.vcf > ${outdir}/${sample}_genotyping_biallelic.stats

done

eblerjana / pangenie

Issue with VCF (does not represent a pangenome graph) #82