joint variant calling VS single sample variant calling
I did single sample variant calling with samples from a trio before. Now we used the same version of GATK and did joint variant calling on three samples together (only run UnifiedGenotyper step, using the old bam file). The command is almost identical except inputing three bam files. After compare the results, we found that all the differences are variants that are either new or not called. No variant has genotype changed. Do you think this is possible? Is that because the genotyping is still based on individual data.