As a reminder, this is part II of my discussion with the learned Mike Magoon, who runs the blog https://open.substack.com/pub/frompovertytoprogress
Here was the discussion we had:
But there are also organizational and societal effects. I have read numerous media reports about rapid growth of DEI in medical profession, including:
1) Forced hiring and admissions to medical schools of sub-par individuals because of their race.
2) Prohibitions of genetic research that might in any way show genetic differences between racial and gender groups.
3) Rejection of the concept of biological gender and pushing “gender-affirming” care on youths.
My discussion of race and medical school admissions (issue 1) is here
So let’s talk about research rooted in genetic differences between racial groups.
Racially narrowed research is quite valuable. There are a number of diseases that afflict some racial groups more than others, either with increased frequency or with worsened severity. On trial in particular stands out: the Blacks with Heart Failure trial. The trial randomized 1050 self-identified African American patients with advanced congestive heart failure to either: conventional therapy or conventional therapy plus the combination of Isordil+Hydralazine (formulated as BiDil). The trial was terminated early because the BiDil+conventional treatment arm had a 6% mortality after 1 year, compared to 10% mortality in the conventional treatment arm.
The trial was and is very valuable, although it is not clear how widely prescribed this drug is in African-American patients with congestive heart failure. This paper suggests that only a small fraction of African American patients receive a prescription for BiDil, based on a drug sales volumes. The authors estimate that 750,000 African Americans suffer from CHF. This would imply a sales volume of $500 million-1 billion. In reality, sales volumes run around $12-15 million dollars. At this point, even a major push to prescribe BiDil is unlikely to make a great deal of money for the company that conducted the research (NitroMed, Lexington MA). There are a number of reasons for this low sales volume. Some possibilities:
It may be the case that most patients with CHF get good relief with conventional treatment and don’t want to add on another medication.
It may be the case that this regimen is not so easy to tolerate, in conjunction with conventional treatment
Both Isordil and Hydralazine are available as generic drugs, and patients may be getting the drug separately as opposed to a combination pill
Clinicians may either be unaware of the data or skeptical of the data, and do not make prescriptions for this regimen
Most likely, all the explanations have some plausibility, which implies that BiDil is under-prescribed in African American patients. This makes future investments in racially stratified research less appealing going forward. Still, there are clear efforts at defining interventions that improve health outcomes in African American patients.
This trial enrolled patients who self-identified as African American. This is probably a fairly accurate way to identify people with predominantly African ancestry, but there are some caveats. First, perhaps 20% of African Americans may have been able to “pass for white”. This means that the prevalence of people with African ancestry may be higher than suspected and might not be captured by self-identification. Secondly, ancestry and race are not necessarily coterminous. Ancestry is traced typically by looking for sequences of DNA that are quite rare, and can be traced back to original ancestors from distant places. However the presence of these rare stretches usually correlate to extremely local areas and not whole continents. The difference in genetics between different parts of the continent are very large. In the diaspora, on the other hand, there is often at least some intermarriage in the intervening generations. So while there may be some superficial resemblance between people from the diaspora and people from the continent, there are also significant difference, making genetic conclusions difficult to draw. Self identification may not
Industrial scale genetic sequencing means that we can now collect almost comprehensive sequencing data from people. Is it possible to characterize the race of people without depending on self identification? Race is complicated, and not easily reducible to ancestry or physiognomy. However, it might be possible to predict skin color from genetics, without depending on self identification. This study identified 37 genetic loci that had single nucleotide variations. Different combinations of these variations was extremely predictive of skin color.
Interestingly, this study was sponsored by the National Institute of Justice, so predominantly was optimized for law enforcement. Perhaps this may need modifications for biological or clinical studies. So it is possible in theory to make good educated guesses about skin color based on DNA sequence, and from there, to make pretty good guesses about race. It won’t be perfect, but probably no strategy will perfectly identify race.
But what about the second half of the question? We may be able to do genetic sequencing to identify race. How will we do genetic sequencing to identify the genetic basis of traits, such as height, propensity for illness, or responsiveness to medications?
This is where genetics research gets almost impossibly hard. Most traits, including skin color, exist on a spectrum, which often imply numerous locations (loci) on the genome, contributing to polygenic inheritance. A skin color score, with gross stratification into 5 strata (very pale, pale, intermediate, dark, very dark) depended on 37 loci.
This study below, from the UK, identified several polygenic scores (PRS) that modestly correlated with height. The best score, PRS.2, required 30,000 variants to show a modest correlation with height.
How in the world can a score like this lead to any kind of intervention? We cannot possibly edit thousands of locations into favorable sequences. We cannot possibly even guide assortative mating, beyond the most obvious (marry a tall person if you want a tall child). Lastly, how can these innumerable locations give us any insight into the mechanism by how genetic sequence predicts or determines height?
Beyond the underlying sequence, genetics is plagued by two different kinds of “dark matter”. On the one hand, epigenetic modification of the DNA vastly complicates a hideously complicated picture. On the other hand, we have the vast, vast tracks of non-coding DNA, that will never get translated into protein, and we have no idea why these stretches have been lovingly maintained for billions of years. Epigenetics reflects the modification of the DNA with “add-on” molecules, most commonly methyl groups. These additions are added to roughly 1.5% of all the bases in the DNA, which means perhaps 90 million bases, and the add-ons shift over time. Furthermore, even in the same person, at the same time, the epigenetic modifications can be widely different in white blood cells as opposed to fat cells.
A fascinating study from France looked at epigenetic modifications following weight loss surgery. Some 46 areas of white blood cell DNA had changes in their epigenetic markings after weight loss surgery. Most of these areas were in long stretches of DNA that don’t participate in protein production. So what are those areas of DNA doing? Why are they getting methylated or de-methylated? Are the epigenetic changes following weight loss surgery in white blood cells even have any relevance, mechanistically, for weight loss? Is this even the right tissue to study? How will this contribute to designing a generation of drugs or other interventions that might be less invasive than weight loss surgery? These are all difficult questions, and I don’t think that adding the complexity of racial markers will help much in teasing this out.
The most obvious point here is that any remotely trained biochemist or geneticist knows all of this. She would know these and innumerable more studies, and beyond the studies, she would understand the underlying complexities. It is insanely difficult to find a genetic basis of any trait. Identifying the mechanism of these genetic sequences, most of which are non-coding, and highly regulated by epigenetics, is even more maddening. Why would she add race to an already difficult picture? More importantly, why would any funding entity be interested in funding what is almost certainly a wild goose chase? And this is just for something like height, which is easily measurable. Intelligence, temperament, moral character would be even more difficult.
If she were truly hellbent on finding a genetic basis for intelligence, temperament, moral character and so forth, and then correlating it to race, her best bet would be to hit up some conservative billionaires, and setting up her own institute, fully equipped and staffed with private money, to chase this mirage. Of course, any billionaire worth their billions would be very wary of writing checks for charlatans.
In general, I will try to provide solutions and my own ideas instead of tearing other ideas down.
‘How would I proceed with genetic research for traits?
I would start by looking for incredible outliers, who have much more normal siblings, parents and children and then look for differences in the genetics. This simplifies some of the challenges of nurture, with presumably similar epigenetics. It also simplifies the broad genetic background because most of these relatives will have similar genetics.
The French study looking at epigenetic changes after weight loss surgery is terrific. More of this, please.
I find myself drawn to the possibility that non-coding DNA may be the source of RNA stretches that might be shepherding pathways of enzymes into the same space.
It may also be the case that these stretches of DNA may be coordinating these enzyme complexes. The PhD thesis of Krishna Shrinivas, charmingly entitled “Dewdrops on the genome” describes some evidence for how sequence drives function.