SNP calling reads number

[wanguan2000]

I followed the GATK pipeline.
when I use the IGV to see the bam file I found the question:
position chr17:7578524 : G(51 reads)-A(20 reads) I think this is a snp, but the position doesn't exist the result VCF files. GATK don't think this is a snp,even in lowQual snp.
why? the position have so high sequencing deepth.Who can explain it

[Robby]

That could have many reasons. How are the mapping qualities of the reads at that position and how are the quality scores of the bp supporting the SNP?