[Vcftools-help] missing data and depth of coverage

[Billie G. <bil...@ut...>]

Hello all, 
I have a hopefully not too naive question about missing data and depth filtering with vcftools.

When I filter my vcf file so that it has no missing data for any individual (using option --geno 1) as I understand it this should eliminate any marker with low or no read coverage in any one individual.

command:
vcftools --vcf good_variants_final_autosomes.recode.vcf --maf 0.0625 --geno 1 --out Variants_screened --recode 


But when I then calculate per site total depth for each marker in the file (using --site-depth output option) . . .

vcftools --vcf Variants_screened.recode.vcf --site-depth

I find that the minimum site coverage value (total across all individuals) is 1. This means there is one or more sites with a total of 1 read covering that variant site (across all individuals (n=32)) in the vcf file. 

How is this possible? shouldn't the first missing data filter have eliminated these variants from the file ?

any advice much appreciated!