Gholson Lyon calls for more rigorous standards in genetics research so that results can be disclosed to subjects and their families as valid clinical diagnostic information (Nature 482, 300–301; 2012). This well-intentioned proposal is impractical because it would add huge costs for only occasional benefit.

Lyon is essentially recommending that the standards of Good Clinical Practice (see go.nature.com/1pado7) should apply to laboratory research as well. The cost and complexity of carrying out all research to comply with these regulations would be crippling. Compliance would demand validated environments for sample-tracking (Lyon's focus) and for all analytical work. In whole-genome sequencing, validation for error-free determination of those 3 billion or so data points in the human genome would be almost impossible.

A more pragmatic alternative would be to incorporate specific provisions in the informed consent for people wanting to access their data. This would carry the disclaimer that results are not formally validated for clinical use, which would require independent replication to meet the standards of Good Clinical Practice. This re-testing could be made available at no cost to the patient should the clinical need arise.

The cost of responding to such an occasional requirement would be insignificant in comparison with processing every subject's specimens under regulatory-compliant conditions.