Discrepant MAPQ and NM values in BAM and Vardict output

[jkunisak]

Hello,
I am running VarDict (v1.8.2) with the below parameters, followed by a var2vcf_valid.pl to convert the output to VCF.

VarDict command:
vardict -G $genome.fasta -p -f 0.0 -N $samplename -b $bamfile -y --nosv -T 0 -m 8 -Q 0 -r 0 -c 1 -S 2 -E 3 -g 4 $intervals_bed_file > $vardict_output.txt

Var2Vcf command:
var2vcf_valid.pl -a -A -c 0 -m 8.0 -Q 10.0 -f 0.0 -P 0 -v 0 -E -N $samplename $vardict_output.txt | gzip > $vardict_output.vcf.gz

At one of the positions, chr1:45331833, there is a C>G mutation called with the following VarDict filters v2;Q10.0;SN1.5;NM8.0. This mutation is supported by 1 read so the v2 filter makes sense.

Looking at the BAM file and grepping for the read supporting the mutation, the mismatch count is 1 (for the mutation itself), rather than 8. Additionally, the read mapping quality is 60, rather than being <10. The same reference genome (hg38) was used for the alignment and variant calling step.

Do you have any suggestions on how I can resolve this? Thank you.