About QC metrics during gVCF merging

[taejun12]

Thank you for developing such a useful tool.

I have several questions about the method used to merge gVCFs.

I merged the gVCFs of multiple samples generated by DeepTrio using GLnexus. I understand that, in a gVCF's non-variant site block from DeepTrio, DP and GQ represent the minimum value within the block, while PL shows the first value.

chr1    594888  .       G       <*>     0       .       END=594954      GT:GQ:MIN_DP:PL 0/0:50:19:0,60,599

If both chr1:594888 and chr1:594950 are included in the pVCF, how are their GT, GQ, DP, AD, and PL values determined?

Will the GT, GQ, DP, and PL values for chr1:594888 and chr1:594950 both be 0/0:50:19:0,60,599 as shown above? And how is the AD value decided?

Also, in multi-allelic sites in gVCF, how are AD and PL values determined?
For example, in the gVCF:

chr1    407272  .       CCACCAACCCCGGGAACCGCCTCCCACTCCCCCCACCAACCCCGGGAACCGCCTCCCACTCCCCCCACCAACCCCGGGAACCGCCTCCCACTCCCCCCGCAACCCCGGGAACCGCCTCCCACTCCCCGCGCAACCCCGGGAACCGCCTCCCACTCCCCCCACCAACCCCGGGAACCGCCTCCCACTCCCCCCACCAACCCCGGGAACCGCCTCCCACTCCCCCCACCAACCCCGGGAACCGCCTCCCACTCCCCCCACCAACCCGGGAACCGCCTCCCACTCCCCCCACCAACCCCGGGAACCGCCTGCCCCCGCCACCAACCCCGGGAACCGCCTCCCACTCCCCCCG    C,<*>   0RefCall .       GT:GQ:DP:AD:VAF:PL      0/0:33:18:16,2,0:0.111111,0:0,33,47,990,990,990

And in the corresponding pVCF:

chr1    407461  chr1_407461_CCCA_C;chr1_407461_C_T      CCCA    C,TCCA  47      .       AF=0.375,0.125;AQ=16,47 GT:DP:AD:GQ:PL:RNC      0/2:13:7,0,6:47:47,990,990,0,990,53:..  1/1:11:3,8,0:6:8,6,0,990,990,990:..      0/1:12:5,7,0:16:16,0,43,990,990,990:..  0/0:18:16,0,0:33:0,0,0,0,0,0:..

For the sample in pVCF with 0/0:18:16,0,0:33:0,0,0,0,0,0:.., how are these values derived from the original gVCF?