Clustering a rare variant on the GDA array

Question

[panacotaforcota]

Dear All,

I work with genotyped data from the Infinium Global Diversity Array (GDA). In GDA array there are variants with known clinical significance identified from clinical WGS and WES samples (by American College of Medical Genetics (ACMG) Recommendations for Reporting of Secondary Findings.

I know that the quality of genotyping of rare variants leaves much to be desired due to problems with clustering. But I would like to understand how to detect these problems using some metrics.

For example, in the figure there is clustering of a very rare variant rs886040456.

There cannot be such a number of heterozygotes. But all metrics are normal:

Do you have any ideas why there might be such a picture?

Thanks!