For the last six months, I've been focused full-time on my health. I've spent tens of thousands of dollars on blood tests and doctors to discover that I have several allergies, I'm vitamin D deficient, and I'm at risk for cardiovascular disease.
In other words, I'm an Indian man.
Discussing genetic differences between racial groups has been stigmatized and for good reason. In the past it’s been used as a tool for genocide and segregation. But with the rise of countless longevity startups and self-experimentation (e.g. Bryan Johnson), we're missing a critical part of the equation: race. Because race is so controversial, I'll try to stick to examples involving Indians in this post.
Indians are very likely to be lactose intolerant; East Asians often have difficulty metabolizing alcohol (ALDH2 deficiency); Indians metabolize more carbs than fats compared to other races. And genetic differences go beyond just race. Women tear their ACLs at much higher rates than men, because their pelvises are shaped for birthing.
But let's go even further. Is there a genetic reason why Indians smell bad? Were certain races more resistant to COVID-19? Are there important differences in how each gender learns?
Nutrition science & pedagogy are widely accepted to be bullshit, and this is a major reason why. The research and advice pays little attention to our physiology, which can greatly vary based on genetics. This is important to make the most of our bodies and minds.
So I want to propose a new-and-improved discipline: Race Science 2.0 holistically looks at history, biology, culture, and psychology to explain traits based on ancestral lineages, enabling individuals to command better control of themselves.
Real race science has never been tried.
The original race "science" had a glaring issue: it was incredibly unscientific. Phrenology (skull measurements) and graphology (handwriting) have no relationship to IQ, and these practices were promoted to justify discrimination and segregation in the era of Jim Crow.
Race scientists need to be the most intellectually honest, because there is so much that can go wrong:
Source: eigenrobot
Today, we are still relatively unscientific in our approach. It is widely cited that at most 10-15% of genetic variation can be explained by race. Yet this is based on an oddly racist definition of race. Even the casual racist knows that you can't group all of Asia or Africa into a single race.
The trick is to double down and recognize the multiple races within each continent (e.g. India is obviously distinct from broader Asia), where there is clear heritability for traits like intelligence, attractiveness, height, and cancer risk. After this refined clustering, what falls under this 10-15% (probably more), and what does not?
There are existing examples of "good" race science. The controversial paper Natural History of Ashkenazi Intelligence" looks at biology, history, and culture to explain why Ashkenazi Jews are intelligent. Another example: it was discovered that the ABCC11 gene, which causes East Asians to have dry earwax, also makes their sweat less pungent. To my knowledge, it has not been seriously investigated whether there is a comparable gene that causes Indians to smell worse.
Bad race science is almost everywhere that race is mentioned. For example, in this explanation for why Indians are vitamin D deficient, nowhere is it mentioned that Indians biologically need more sun exposure for an equivalent amount of vitamin D production. Instead this is chalked up to purely cultural differences.
Modern science does an acceptable job of studying questions related to nurture:
- Are there environmental factors that cause or exacerbate the undesirable phenotype?
- Is there a stated reason for the trait according to the relevant population?
I am interested in additionally studying the following:
- What is the reason for the presence of this trait in the ancestral populations, if such an explanation exists?
For example, are there certain regions of India where hygiene is paramount to preventing disease, and therefore these populations adopted a worse smell to signal when someone is not hygienic?
While a complete genomic sampling for an individual is the gold standard, we might first look to their familial history, (critically) their race, shared traits across all humans, and evolutionarily conserved traits from our pre-humanoid ancestors.
- If applicable, what is the specific genetic mutation responsible for this phenotype?
This might often be very difficult to determine. You can prove heritability via twin studies or looking at adopted children, but these cases are rare. You can also never fully control such an experiment. But we can certainly identify differences in average physiological makeup between genetic clusters.
- Possibly most interesting for race science: What can we do to mitigate or reverse the undesirable effect?
How is this actionable?
- Precisely define each race, likely via genetic clustering.
The clusters should not (necessarily) reflect continents, countries, or any existing group. But we do expect the clusters to be phenotypically similar.
- Define the methodology for approaching questions in the field.
Race Science 2.0 aims to make predictions about individuals, not groups, so population-level R2 isn't as important as individual-level R2. What other constraints should we impose?
- Establish consensus on ethical guidelines and in what areas the data is actionable.
This approach has clear applications in nutrition and longevity. What else?
This field might find difficulty publishing its results in legacy scientific journals or even getting approvals from institutional review boards initially. It motivates the need for a revamped scientific enterprise, alongside a privatized "FDA" which can independently evaluate the science that this enterprise produces. Such an entity is useful beyond race science. For example, the entity ought to be equipped to interpret early bleeding-edge results that might conflict with scientific consensus, such as data from Bryan Johnson's self-experimentation, physicians who manage to cure their own cancer, and mRNA vaccines. It might forego the traditional three-phase clinical trial requirement if strong statistical evidence can otherwise be provided.
I would love to connect with others who are interested in this area. If you're like me, you've identified some areas where AI is trained to self-censor or give politically acceptable outputs. For some period, human intuition might reign supreme in race science, so there is plenty of work to do.
