Thread by @AdamRutherford

CRISPR gene editing for intelligence. I am writing about this idea in my new book on eugenics, which right wing provocateurs seems to be chuntering about these days. Some brief, practical thoughts:

https://twitter.com/charlesmurray/status/1379114335641993216

Set aside the ethical issues, and just focus on the scientific and legal for now.

Legally, germline editing in humans is internationally prohibited in human embryos that can come to term (14 days post fertilisation is typical and heavily regulated limit).

(Mitochondrial replacement gene therapy is one exception, though this is not gene editing so much as replacing a whole chromosome.)

He Jiankui (and maybe others unknown) criminally broke those laws, but in doing so also did not successfully demonstrate that it was possible, practical or feasible.

His proposed edits (CCR5 32) were *not* implemented in the two embryos that were experimented on, and subsequently re-implanted and were born. Instead, two new alleles were introduced, unknown to nature or science.

Furthermore, off target mutations were not (to our knowledge) assessed. We also don’t really know how the 32 mutation works in preventing HIV infection, and it may be detrimental for other disease modalities, such as West Nile virus.

In two earlier and legal experiments on correcting monogenic disorders with CRISPR (Sudden cardiac death and -thalassemia) resulted in >20% of cells in a blastocyst uptaking the corrective edit (and from memory, off target mutations were not assessed).

{gah, that's delta-32 and beta thalassemia}

What this means is that the edit efficiency is low, and we have no way for controlling the tissues in which the edit will end up – the cells in the blastocyst will end up as placenta as well as all over the body, but not necessarily in the affected tissue (e.g. the heart).

Now, we have a far faaaar greater understanding of monogenic traits than for complex traits, and none are more complex that intelligence or cognitive behaviours.

The alleles identified so far that relate to intelligence, IQ or cognitive behaviours are very poorly understood, are expressed in multiple tissues (not just the brain) and have small effects.

As for the phenotype, IQ is a valid metric for intelligence, fluctuates during life, and is better understood at a population level rather than for individuals. The genetics of IQ are poorly understood, extremely polygenic, and the effect size for specific SNPs is mostly unknown.

So, in summary, germline gene editing in humans is legally prohibited. It is scientifically *possible*, but realistically only theoretically atm, and attempts so far have been necessarily simplified, and are still riven with fundamental problems.

GATTACA was a fiction, and shouldn't be taken as gospel for what is feasible or possible.

I, as someone who knows a little about genetics, think that discussions about gene editing for intelligence as a commercial prospect are comically premature.

...which doesn't mean we shouldn't think about it.
But, *if* possible/successful the phenotypic marginal gains would probably be swamped by environmental constraints, and would take years to be seen, if at all.

It would be cheaper and more efficient to improve literacy or numeracy by educational means.

And then there’s the ethical issues. But that’s for another time.

Comments from @ewanbirney worth reading here too, e.g.

https://twitter.com/ewanbirney/status/1379392726446309377