Clair3 GVCF created file is 1bp of the end sequence so validatevariants fails

[klausyboi]

Hi,

Im trying to combine my gvcfs, first using gatks validatevariants but i get the error:

A GVCF must cover the entire region. Found 16 loci with no VariantContext covering it. The first uncovered segment is:Pf3D7_01_v3:640851

Looking at a sample that worked from illumina the bp is always 1 short for each chromosome and 640851 is the end of the chromosome in this case so i think thats the problem.

I used clair3 to get the gvcf using: barcode=\$(basename "\$barcode_dir") run_clair3.sh \ --bamfn=barcodes/\$barcode/calls"\$barcode"_sorted.bam \ --ref_fn=${ref_seq} \ --threads=${threads} \ --platform="ont" \ --model_path=/mnt/storageG1/data/software/rerio/clair3_models/r1041e82${bps}_${type}_v500 \ --output=vcf/\$barcode/ \ --include_all_ctgs \ --gvcf

I then used sortvcf from gatk because the variants were all in the wrong order and I was getting this error, hoping it would fix it but alas it did not.

Any help with this is greatly appreciated.

Thanks

[aquaskyline]

Could you please send over the reads and reference sequence of Pf3D7_01_v3. We will try the repeat the problem on our side.

[klausyboi]

ref genome my bam my fastq if you wanted to start from there which was created from the bam anyway but otherwise id be sending over lots of pod5s

[zhengzhenxian]

@klausyboi Thanks for sharing the data; we will fix the gvcf range issue in the next release.