If you get your genome processed into a VCF file, each variant will be listed together with sort of a posterior probability. The discrepancy will be proportional to how stringent your filter is, as variant calling methods are well calibrated. There's a lot of literature on this.