There’s plenty of genetic differences that map, roughly, to ethnic lines. Lactase persistence, odds of getting a stroke, sickle cell, and of course the obvious ones like hair type, and hair and skin color.
With intelligence tests, there’s a clear incentive to lie to make your own “race” look better*, there’s so many interconnected variables that there’s plenty of ways to hide bias in all the complexity, it’s not at all obvious how to separate environmental influence from genetic influence**, and baseline intelligence might be overrated anyway compared to things like willingness-to-learn. Intelligence testing is at least as complex as athletic testing, and nobody’s going to deny that flexibility, stamina, and power are all different things, and that a great swimmer, a great boxer, and a great dancer will have different body types, in spite of them all being athletic powerhouses. This means it is completely possible for there to be real genetic differences in intelligence, but a single score like IQ is insufficient to describe what they even are.
This all makes intelligence testing a wicked problem. There’s definitely a problem, but there is no consensus on what the problem really is, there are incentives to sabotage any solution, there is no real way to tell if it’s solved or even getting better, and it’s subject to the observer-expectancy effect: any attempt to measure intelligence will cause people to train themselves to do better on the test (even if there is no incentive other than bragging rights attached to it), which means measuring the system also changes the system.
I think I have extremely good reason to be skeptical claiming to have isolated any particular variable influencing intelligence.
* Yes, I know about the dairy industry quietly covering up how common lactose intolerance is. You get a cookie. And a glass of milk to go with it.
** Racism is an environmental factor, not genetic, yet there are obvious reasons why identical twin studies won’t fix that variable.