Filtering variants on the basis of average depth plus 3 standard deviation

Question

[nadiabeg.comsats]

Hi all. I am working on variant filtration for my vcf file and found following parameters in one paper . To get high quality SNP they filtered variants on 1) raw read depth (DP) average sequence depth multiplied by the number of accessions; and 2) DP lower than average DP plus three standard deviations (SDs). I am unable to understand these parameters. Anyone knows how to do these calculations practically. According to my understanding It must be like" I should calculate the mean first for all the data points (here data points are DP column in my vcf file), then variance and then SD*3. I am not sure what they mean by average DP here.