What a genotyping array can tell a laboratory, and what it can’t.

What an array measures

A genotyping array measures a preselected set of single positions in the genome, usually several hundred thousand to a few million, chosen when the chip was designed. At each of those positions it reports which bases are present. It does not read a gene from end to end.

Where arrays are strong

Arrays are reliable and inexpensive for common variants that the chip was designed to measure. Many well-studied pharmacogenomic markers and common risk variants fall into this group, provided the chip includes them. Whether it does depends on the assay, so it is the first thing an evaluation checks.

What arrays do not see

Most deletions, duplications and other copy-number changes are invisible to an array, and so are repeat expansions. So are the many rare variants that are simply not on the chip. For genes where disease is usually caused by changes like these, a result with no variant found is not a clear screen.

Rare calls need confirmation

The rarer a variant, the less reliable an array call. A 2021 study in The BMJ by Weedon and colleagues compared array results with sequencing and found that, for very rare variants, most positive calls were not real. A rare pathogenic finding from an array should be confirmed with a validated clinical test before anyone acts on it.

What this means for a report

A report built on array data has to show its limits. In an Atlagene report, every category states how much of its panel the file measured. Positions the file never measured are left out of the result and are never presented as normal. The laboratory decides, with its own validation, which results are fit for clinical use.