Skepticism about polygene scores to select for IQ and height

One caution when objecting to the prospect of heritable human gene editing is to take care not to overestimate what it technically possible.  That is, an all-too-easy argument is that attempts to edit a disease gene will lead, by momentum if nothing else, to “designer babies,” with children not just being genetically selected but in fact engineered in great detail for traits like attractiveness, athletic prowess, height, and intelligence.  This contributor to this blog has repeatedly taken the position that heritable human gene editing is a project that fundamentally alters the way we see ourselves and each other; that divides the human race into “actors” and “acted upons;” that has no prospect of prospectively assessing long-term, unintended consequences, to an individual subject, subsequent generations, or society at large; and that fortifies a perspective of admitting to the human race only those members we want to admit.

Along the way, we must keep in mind that “designer babies” are not likely to be feasible in the foreseeable future.  One recently-reported case in point is a study by scientists at the Hebrew University of Jerusalem.  A preprint (in advance of publication in a peer-reviewed journal, it is said) is publicly available here.  I daresay the details will be inaccessible to all but specialists in genetics, but a summary of key points is provided by a technical writer at a website called GenomeWeb.  In brief, some of those points:

  • A score based on assessment of multiple genes has previously been suggested to explain only about 5% of the difference between individuals in IQ (300,000 people genetically tested) or 25% in height (700,000 people tested).
  • These researchers tested about 1000 people, and considered about 15,000 genetic variations.
  • They looked at offspring of actual couples and also “simulated” matches for about 500 would-be couples made from individuals for whom they had genomic data.
  • Of note, they appear to have looked at “SNPs,” or “single nucleotide polymorphisms,” which are relatively easy to catalog across the 30,000 or so human genes, and which themselves run into the hundreds of thousands across those genes, but SNPs are far from the whole genetic story.  Larger differences in genes, or how those genes are translated into biological traits, is much more complex to assess.
  • They surmised that, if their score were used to try to predict height, the average gain would be about 2.5 cm (about one inch), with a range of 1-6 cm.  If used to predict IQ, the average gain would be about 2.5 points, with a range of 1-7 points.
  • Then they also looked at 28 actual families with lots of kids, from 3 to 20 (!).
  • For the actual families, the score predicted to cause the tallest child did so for only 7 of the 28 families, and the highest scoring child was actually shorter than average in the family in 5 of the 28 families.  No attempt to assess IQ for these real families, apparently.
  • They point out other reasons why trying to select for IQ might be problematic—potential association with autism and anorexia, for example, as well as just general complexity.
  • They suggest that for most people undergoing IVF, and creating fewer than 10 embryos in the process with less than 100% success after implantation in the womb, the odds are not good for making a reliable forecast of an offspring’s height or IQ.
  • They make these points without commenting more broadly on the ethics or policy wisdom of allowing or encouraging heritable genome editing to proceed.

A complex story, and a developing one, to be sure, but one should not be too quick to accept grandiose promises for predicting complex traits based on genetics.  At least for now, those appear to be rather “ahead of the puck,” shall we say.

Leave a Reply

Please Login to comment
  Subscribe  
Notify of