A question about using GenomeAnalysisTK to deal with ancient DNA data

Hi,Team,
I want to call variants for some ancient samples using HaplotypeCaller.
My command is as follows:
java -jar GenomeAnalysisTK.jar -R myref.fa -T HaplotypeCaller -ERC BP_RESOLUTION -I 1.bam -L site.list -o 1.genotype.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T HaplotypeCaller -ERC BP_RESOLUTION -I 2.bam -L site.list -o 2.genotype.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T HaplotypeCaller -ERC BP_RESOLUTION -I 3.bam -L site.list -o 3.genotype.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T CombineGVCFs --variant 1.genotype.g.vcf.gz --variant 2.genotype.g.vcf.gz --variant 3.genotype.g.vcf.gz -o merge.g.vcf.gz
java -jar GenomeAnalysisTK.jar -R myref.fa -T GenotypeGVCFs --variant merge.g.vcf.gz --includeNonVariantSites -stand_call_conf 30 -stand_emit_conf 30 -o merge.vcf.gz
In the 3.genotype.g.vcf.gz ,a site is :
1 72710 . G C, 2.99 . DP=1;ExcessHet=3.0103;MLEAC=1,0;MLEAF=0.500,0.00;RAW_MQ=3600.00 GT:AD:DP:GQ:PL:SB 1/1:0,1,0:1:3:25,3,0,25,3,25:0,0,1,0
In the merge.g.vcf.gz ,the site is :
1 72710 . G C, . . DP=1;ExcessHet=3.01;RAW_MQ=3600.00 GT:AD:DP:GQ:PL:SB ./.:0,0,0:0:0:0,0,0,0,0,0 ./.:0,0,0:0:0:0,0,0,0,0,0 ./.:0,1,0:1:3:25,3,0,25,3,25:0,0,1,0
But in the merge.vcf.gz ,the site is :
1 72710 . G C . . AC=0;AF=0.00;AN=2;DP=1;ExcessHet=3.01;MQ=60.00 GT:AD:DP:RGQ ./.:0,0:0:0 ./.:0,0:0:0 0/0:0,1:1:3
Why the site with low QUAL changes from 1/1 to 0/0 finally but doesn't delete it ?
What kind of impact is the result of this for me ?
Because my variantsite is few,every site is important.
Could you suggest how to fix this problem?
Thanks!
best,
KuroKami