PGT only available for some variants in GATK .vcf

Question

[michael.flower.14]

I've got a vcf file someone else prepared using GATK. I'm interested in the phasing information in the PGT tag e.g. 0|1.

This information seems to be available for some variants, but not for others e.g. below

chr1    16977   .   G   A   92.77   PASS    AC=1;AF=0.500;AN=2;BaseQRankSum=-3.368e+00;ClippingRankSum=0.00;DP=24;ExcessHet=3.0103;FS=0.000;MLEAC=1;MLEAF=0.500;MQ=22.97;MQRankSum=-1.010e-01;QD=3.87;ReadPosRankSum=-7.950e-01;SOR=0.818;VQSLOD=1.74;culprit=QD    GT:AD:DP:GQ:PL  0/1:17,7:24:99:121,0,437
chr1    19391   .   G   A   152.77  PASS    AC=1;AF=0.500;AN=2;BaseQRankSum=-7.830e-01;ClippingRankSum=0.00;DP=22;ExcessHet=3.0103;FS=0.000;MLEAC=1;MLEAF=0.500;MQ=21.83;MQRankSum=0.825;QD=6.94;ReadPosRankSum=-3.500e-02;SOR=0.760;VQSLOD=2.62;culprit=QD GT:AD:DP:GQ:PGT:PID:PL  0/1:15,7:22:99:0|1:19355_G_A:181,0,612

I've subsetted two SNPs I'm interested in using the following script, but unfortunately both are SNPs without PGT information.

vcftools --gzvcf "$VCF" --positions "$DIR"/positions_hg19.txt --recode --recode-INFO-all --out "$DIR"/positions

Anyone know why PGT is missing for some variants?

Answer 1

[Pierre Lindenbaum]

Anyone know why PGT is missing for some variants?

There is no magic: Genotypes can be phased only if GATK finds two variants on the same read fragment.

[michael.flower.14]

Are there any options I can use to improve the phasing and get PGT information for more SNPs?