Substack cometh, and lo it is good. (Pricing)

Ambry Genetics and PRS “equity”

A few people shared with me some screenshots from a message from Ambry Genetics regarding polygenic risk scores (click to enlarge). From what I can gather this seems to indicate that they’ll stop offering polygenic risk scores until they become more equitable by population. The basic issue here is that a prediction for one population may not port over well to another. The standard “solution” to this problem has been to expand population coverage by doing more research on non-European populations.

Perhaps I’m misunderstanding, but the communication here seems to suggest that they will stop giving scores to people of European ancestry until they can provide them to other groups in the interests of equity.

The communication here was shared on genetic counselor Twitter, and the messages were all very supportive. Generally, they were of the form “this is how antiracism is done!”

If I understand this correctly, I am not supportive. The traditional means by which these sorts of inequities were remediated is through addition, not subtraction. Not providing a polygenic risk score that might be actionable and useful to white Americans would be like simply rejecting using standardized tests because there is a gap between races.

There are two major issues here. First, “white Americans” are heterogeneous socioeconomically. White American life expectancy is declining. This decline is driven by social and class stratification. A focus on racial equity masks this reality. Due to the interaction of the environment with genetic risk improving equity through subtraction will increase class inequity. The white Americans who could benefit the most from polygenic risk scores because of all their environmental risks are probably Hillbilly Elegy whites.

Second, there has and will be progress in diversifying the training sets with non-Europeans. We can’t get worse than where we are. The issue is not racism in the classical sense, but various structural and historical forces (outside of East Asia most of the studies have been done in the USA and Europe). And, you can’t chalk up all the blame to Americans and Europeans. India and the Middle East in particular have great possibilities in the domain of biomedical genetics due to inbreeding and endogamy but have not leveraged their natural advantage in this area due to bureaucratic inaction and lack of execution. The scientists in these nations want to “go” but the political and institutional support is not there, or slows them down.

I must also add that I have heard privately from some researchers who are white that they have made a private decision to shy away from research on non-Europeans due to the possibility of controversy. Basically, some activists may not believe it is a “good look” for white people to be doing research on datasets of “people of color,” so on a personal level of decision-making, some white scholars are just leaving the work to nonwhite scholars. This is not a good trend in my opinion, though I understand why individuals with really diverse career options would choose the maximum gain for the minimum risk.

19 thoughts on “Ambry Genetics and PRS “equity”

  1. The traditional means by which these sorts of inequities were remediated is through addition, not subtraction.

    In theory, yes. But leftists have a long history of destroying overall economic value and engaging in a negative sum governance in an attempt to ensure that the individual distribution of the said sum is more “equitable.”

    some activists may not believe it is a “good look” for white people to be doing research on datasets of “people of color,” so on a personal level of decision-making, some white scholars are just leaving the work to nonwhite scholars.

    I’ve long believed that people deserve the government they elect. I suppose now it will be also the case of people deserving the kind of science they support (or not oppose).

    I guess this is how a great civilization dies – through (assisted) suicide.

  2. The overwhelming majority of genome-wide scans have been on
    people of European descent, with the majority of the remainder being East Asians.

    This is not good and needs fixing and there are a number of initiatives which will help.
    But there are a variety of problems.
    1)
    “Advanced countries” are now generating huge datasets. UK Biobank began with 600K contributors
    and very rich phenotyping. Such a study needs a well developed health care infrastructure and highly expert genetics.
    It will not be easy to match this in (say) Sub-Sahara Africa — and for some populations it may never be matched
    because the population is much smaller. In addition funding these studies is expensive and poor countries have
    other priorities.

    2)
    I also have heard of white researchers wanting to avoid research on minority groups. This is unsurprising
    but will not help the minorities that “woke” claim to want to support. In my podcast
    I mentioned that David Reich and I spent a decade mostly studying African-American genetics and we were criticized
    for racism. That has to be discouraging for a young researcher.

    3)
    There are also technical issues that make study of African genomes a little harder. Genome scans rely on “LD” which roughly
    means that nearby SNPs are dependent. But in Africa, with a different genetic history, the dependence drops away faster
    with distance. That means that it is more expensive to get the same genomic coverage in Africa than in Europe and
    means that polygenic risk scores will probably never be equal in quality, even if other issues like data collection
    and phenotyping are of the same standard.

    Pace Ibram Kendi, not all differences between human groups are due to racism.

    Nick Patterson

  3. Rich, white wokes waging a class war on working class whites, using dubious moral takes as weapons.

    “We’re going to close our soup kitchen, because there’s too many poor white people using our services, but not enough blacks. When there are as many hungry blacks as whites we’ll reopen.”

    It’s almost insane.

  4. Is there a group that hates its own poor as much as whites Anglophones do? Seriously I don’t understand, how did this come about? Is it just classism? But they love poor minorities and only seems to have a vendetta against poor whites.

  5. Maybe there really is a tendency to want to return to a feudal-like system where elite status is inherited. The ones in power are feeling a little too cushy to participate in social mobility.

  6. We becoming dumber and dumber. Self inflicted ignorance is a bad thing.

  7. The race-blind cancer risk predictor PRSs are right around the corner. In fact one reason why the European-only predictors are questionable is purely genetic, not social. Basically too many self described White Americans pack enough Non-European ancestry to make their estimates imprecise. But of course there is a stronger pressure to DEI than to improving outcomes for a subset of the White Americans. In the case of cancer risk prediction the pressures align, luckily.

    But not everywhere. Like it has been reported that the leading hospitals are dropping African American factor from kidney disease algorithm, assigning everyone’s ancestry to non-African. This may cause overtreatment, but is thought to be justified by atonement for the sins of racism. Apparently better outcomes vs. DEI don’t align well there?

  8. Very interesting comments from Razib and Nick; on the issue of LD, I thought one of the benefits of including African populations was that because LD is lower, you’d get a more accurate result of the actual causal variant and not something just in equilibrium with it…

    … but I suppose if you’re looking for PRS scores, you don’t actually care that much about whether you’ve found the true causal variant or just something in very high LD with it. That’s more of a benefit when you’re looking to characterize the true variant to develop drugs and things, which would be more important for single variants of large effect, which would generally be much more obvious?

    Also about how PRS may not be possible at all for some populations. I have wondered before on related topics if population admixture might make PRS much harder for more diverse Western countries, and whether declining family size and more older children might make gold-standard sibling within-family controls much harder.

  9. Re: shorter-range LD, it may help finding the causative genetic variants, but only if they are relatively common in Africa, and only when the datasets are appropriately big (it often takes hundreds thousand people to fine tune cancer risk scores). The relatively smaller African cohorts are still useful for discovery but only the most informative markers may be found there. So it’s often more practical to start from the discoveries made in the European cohorts, and sort of retrofit the originally largely European scores to make them work in other ancestries.

  10. Yes, we is gettin’ dumber. How about an edit function after 5 minutes, Razib.

  11. Shout out to Nick Patterson. I really enjoyed your podcast interview with Razib. I copied it to my mathematician son to show him how careers can unfold.

    Drop by more often.

  12. @Walter Sobchak: So to, my son, who is a topologist. What kind of math does yours do?

  13. My son is in an applied math program. I think it is something at the intersection of statistics and graph theory. But, I blew up in 2nd year calculus. After that it was law school.

    I am very proud of him and he makes good money as a consultant to telcoms. But I can understand his work no better than a pig can understand Christmas.

  14. Is there a competitor to Ambry that offers the same analysis? If one company chooses to turn down business, that’s just more market for someone else. If not, there’s the opportunity for someone with biotech skills to do a start-up.

  15. @abelard impute.me will calculate PRS, also from Ambry sequence data (vcfs). So it’s not availability, just their own engagement that they are shutting down.

  16. I realize this comment is really late, so no one may ever see it.

    Nick Patterson? Your name rings a bell, but I’m too lazy to Google. You are a BFD in genomics, right?

    Wouldn’t some of the LD issues in West Africans be solved by using African Americans, who are a hodgepodge of West African genes? Or is mixed worse for LD? At the very minimum, it might remove the requirement to study every ethnicity separately. If they are just too white, what about Caribbean blacks? Researchers did not say, well white Americans are a mix of too many European ethnicities to do PRS on. At least, they did not say that when they were looking for grants.

    Would sequencing solve the LD problem? Either in general or better for blacks? What about sibling PRS scores?

    Lastly, just how non-white are American whites? I mean older, say, pre-1970 ancestry. I have heard 1% black and 4% Native, but I guess those are roughly doubled for people whose ancestors were colonials? As this research is seemingly done largely by people who want to show intermixing, I had assumed these were overestimates? Certainly white Americans do not look very black or Native, but maybe genes with strong influences on appearance were selected against?

    Mitochondrial and Y chromosomes are somewhat interesting, but do not have to match autosomal percentages. Do they, though, for American whites?

  17. Hey Razib, do iPad comments in safari always post as walls of text?

    Trying a new paragraph

  18. The “colorblind” breast cancer PRS is being presented at ASCO in two weeks.
    https://meetinglibrary.asco.org/record/197017/abstract

    Ancestrally unbiased polygenic breast cancer (BC) risk assessment.

    To make a PRS available for all women, we developed and validated a novel global PRS (gPRS) that utilizes individual ancestral genetic composition… using a cohort of 31,126 self-reported African American patients …
    Conclusions:
    The 149-SNP gPRS is validated and calibrated for women of all ancestries. Combined with clinical and biological risk factors, this approach may offer improved risk stratification for all women, regardless of ancestry.
    (The validation cohorts include African American, Hispanic, Asian, and Mixed-Ancestry groups, and it’s highly significant everywhere)

    Several Black researchers are among the authors just in case if it is important to make laypeople more comfortable…

Comments are closed.