atlageneNote
Validating an array-based report before routine use.
An evaluation shows whether a report fits your laboratory. Validation is your laboratory’s own evidence that it can rely on the report, and it is worth planning before the first patient sample.
Start from the assay
List the positions your chip measures and compare them with the variants the report interprets. A report can only be as complete as that overlap, which is why every Atlagene category states how much of its panel the file measured. Check the export as well: the genome build, the strand convention and how missing calls are written.
Run samples with known answers
Use samples whose genotypes are already established by another method. These can be characterised reference materials, such as those from Genome in a Bottle or the CDC’s GeT-RM programme, or your own samples tested by sequencing. Compare the calls position by position, then compare the reports.
Treat rare calls separately
Accuracy on common variants does not carry over to rare ones. A 2021 study in The BMJ found that most very rare positive calls from arrays were not real. Decide in advance that a rare pathogenic finding is confirmed with a validated test that uses a different method before it is released.
Check the report, not only the calls
Correct calls can still produce a misleading report. Check that its limits are stated, that coverage appears beside each result, that unmeasured positions are left out rather than shown as normal, and that safety-critical findings stay in full whatever the report settings. Then ask your reviewers whether they would sign the wording.
Decide who signs off, and when to re-check
Name who reviews and releases reports, and what is recorded when they do. Agree what triggers a re-check: a new chip or software version, a change to the export, or changes to the variant definitions the report relies on. Atlagene keeps every change to a variant definition in its history, so a re-check can see exactly what moved.
Where an evaluation fits
An Atlagene evaluation runs your laboratory’s actual export through the platform and compares the report with a defined review plan. It does not replace your validation, and an evaluation report is not for patient care. It gives you the material to plan validation within your own quality system.