Interpretation of Homozygous and Hemizygous Counts on gnomAD variants

I have been examining the counts of homozygous and hemizygous individuals for variants on the X chromosome in the gnomAD database, and I am encountering some inconsistencies in their interpretation. Specifically, the way these counts are derived and presented on the gnomAD browser seems to vary, and I would appreciate some clarification.

  1. Variant X-151721385-C-T (gnomAD)
    The number of hemizygous individuals is equal to the allele count in males (AC_XY)
  2. Variant X-312262-C-T (gnomAD):
    The number of homozygous individuals is the sum of the number of homozygous XX and homozygous XY (information taken from gnomad.exomes.v4.1.sites.chrX.vcf.bgz gnomad.genomes.v4.1.sites.chrX.vcf.bgz). However, the allele count in males (AC_XY) does not seem to be interpreted as the number of hemizygous individuals directly and from browser number of hemizygous individuals is 0.

Why is there a difference in how hemizygous counts are derived between these two variants? Specifically, why does AC_XY directly translate to hemizygous counts in one case but not in the other?
Understanding these details would greatly help in interpreting the data accurately and utilizing it effectively in research.

Thanks in advance,

X-312262 is located in the pseudoautosomal region 1 (PAR1) of X and Y chromosome, so it’s treated like in an autosome (no difference for homozygous calling in XX or XY); and X-151721385 is close but outside PAR2, so it’s only present in X chromosome (hence hemizygous calling in XY male).