Skip to main content
Download PDF

No evidence that ACE2 or TMPRSS2 drive population disparity in COVID risks

BMC Medicine volume 22, Article number: 337 (2024) Cite this article

Abstract

Early in the SARS-CoV2 pandemic, in this journal, Hou et al. (BMC Med 18:216, 2020) interpreted public genotype data, run through functional prediction tools, as suggesting that members of particular human populations carry potentially COVID-risk-increasing variants in genes ACE2 and TMPRSS2 far more often than do members of other populations. Beyond resting on predictions rather than clinical outcomes, and focusing on variants too rare to typify population members even jointly, their claim mistook a well known artifact (that large samples reveal more of a population’s variants than do small samples) as if showing real and congruent population differences for the two genes, rather than lopsided population sampling in their shared source data. We explain that artifact, and contrast it with empirical findings, now ample, that other loci shape personal COVID risks far more significantly than do ACE2 and TMPRSS2—and that variation in ACE2 and TMPRSS2 per se unlikely exacerbates any net population disparity in the effects of such more risk-informative loci.

Peer Review reports

Background

In mid-2020, concurrent with early empirical inquiry into roles of host genomic variation in SARS-CoV2 infection [1,2,3,4,5], Hou et al. set out to offer predictive guidance for such efforts, by assessing pre-pandemic public DNA data from two human genes, ACE2 and TMPRSS2, whose protein products were known to interact with other coronaviruses [6].

Pooling public genotype data sampled from various human groups, without phenotypes, they shortlisted ACE2 and TMPRSS2 variants that some computational heuristics predicted likely to alter protein function, and found that most such variants (each, typically, very rare) came from subsets of data labeled “African/African-American” or “Non-Finnish European” versus labeled otherwise (e.g., “East Asian”).Footnote 1 This, they held, suggested real-world population disparities in ACE2 and TMPRSS2 functional variant load, similar for both genes, that might in turn drive population differences in COVID outcomes.

Errant interpretation of genotype data

Alas, Hou et al. had neglected a basic feature of the public data they used—lopsided population sample sizes—that made their summary findings artifactually likely even with no difference between real populations. Specifically, they had pooled genotypes from > 36,000 “non-Finnish European” and > 23,000 “African/African-American” people, but far fewer “Amish” (450), “Ashkenazi” (1662), “East Asian” (1567), or other (< 15,000) people.Footnote 2 As such, even if variants were uniformly distributed across real populations, Hou et al. would likelier find a given rare variant in either of their big samples (“African/African-American” or “non-Finnish European”) than in any of their much smaller samples of other groups.

Consistent with such artifact, the number of ACE2 or TMPRSS2 variants Hou et al. shortlisted for a given population scales well with how many genotypes they sampled from that population (Fig. 1, origin-rooted linear r2 > 0.95 for both genes).Footnote 3,Footnote 4 Long known in theoretical and empirical population genetics [7, 11], the sampling effect apparent in Hou et al.’s summary findings reflects a simple fact: much as counting more of a forest’s birds can help document rare taxa living there, sequencing more of a population’s gene copies helps document rare variants among them. While other factors shape the emergence and fate of such variants, and their rate of discovery with increasing sample size [7,8,9,10, 12, 13], reliably finding and quantifying them entails sampling from many individuals.

Fig. 1
figure 1

Population-specific variant tallies in Hou et al. [6] reflect lopsided sampling. Scatterplots of population-specific tallies (y-axis) of shortlisted variants in ACE2 (orange) or TMPRSS2 (blue), by sample size (x-axis; values denote maximum sampled alleles among shortlisted variant-position genotypes for that gene in gnomAD (v.3.0) + NHLBI-GO ESP6500 genotypes, as pooled by Hou et al. [6]). Datapoints mark values for African/African-American (AFR; TMPRSS2 count excludes 1 variant (p.Pro444Leu) reported by Hou et al., but not in public data, and not consistent with reference variant at given protein residue); Amish (AMI); Ashkenazi (ASH); east Asian (EAS); south Asian (SAS); non-Finnish, non-Amish, non-Ashkenazi European (EUR; TMPRSS2 count excludes 1 variant (p.Gly6Arg) reported by Hou et al., but not in public data, and not consistent with reference variant at given protein residue); Finnish (FIN in Hou et al. [6]); Latino/Admixed American (AMR; ACE2 count includes 2 variants wrongly omitted from this population by Hou et al.); or other (oth; PNA in Hou et al. [6]; ACE2 count excludes 2 variants wrongly tallied in this population by Hou et al.). Best-fit trends (dashed) mark origin-rooted linear regression, conservatively proxying independent (versus cumulative) discovery of potentially selection-constrained (versus selectively neutral) variants in samples from variably sized, mutually diverged populations (versus one steady-sized randomly mating population). We note that even in the contrasting case of cumulative discovery in a steady-sized population, variants under selective constraint (as Hou et al. sought to tally) tend to accrue quasi-linearly, rather than strictly logarithmically, with increasing overall sample size [8,9,10,11]

Full size image

As such, current public genotype data inherently document more of the variants actually circulating in some populations than of those actually circulating in other populations—and allow more precise estimation, in the same best-sampled (if at all) groups, of each variant’s actual frequency (as may figure in functional prediction) or absence. For example, to be 90% confident that even the least rare shortlisted ACE2 variant in non-Finnish, non-Amish, non-Ashkenazi European sample data (p.R219C) is not actually more common in Amish, Ashkenazi, and/or Finnish populations, despite its absence in those groups’ small sample data, would require sampling > 22,700 additional Amish, Ashkenazi, and Finnish copies of ACE2.Footnote 5

Beyond summary tallies, none of Hou et al.’s shortlisted variants reliably proxies any population to begin with. One (TMPRSS2 p.V160M) appears in all studied populations—and in many individuals in each—so offers scant ground to guess which population(s) a carrier comes from. All 130 other shortlisted variants appear too rare in every population to typify members of any of them (even in aggregate, their data suggest that > 96% of people in every studied population likely carry none of those rare variants (Fig. 2)).Footnote 6 And as real populations also harbor unsampled but functionally relevant variants, whose effects on basic protein function (let alone response to a particular virus) current heuristics cannot reliably predict [14,15,16], the tallies and predictions of Hou et al. do not warrant positing that ACE2 or TMPRSS2 (let alone both) functions worse with respect to SARS-CoV2 in any human population (let alone particular ones) than in others.

Fig. 2
figure 2

Nearly everyone, in all studied populations, likely lacks all ostensibly population-distinctive variants shortlisted by Hou et al. [6]. Bar plot of estimated percentage of people in each studied population who likely have none of the 130 notionally population-distinctive (i.e., absent in sample data from at least one studied population) ACE2 and TMPRSS2 variants shortlisted (without empirical evidence for any effect on protein function or other physiology, and omitting many other potentially functionally relevant variants in all populations) by Hou et al. [6]. Estimates (product of binomial probabilities) presume variants assort randomly, independently, at sampled population-specific frequencies, in half-XX/half-XY populaces. AFR = African/African-American; AMI = Amish; ASH = Ashkenazi; EAS = east Asian; SAS = south Asian; EUR = non-Finnish, non-Amish, non-Ashkenazi European; FIN = Finnish; AMR = Latino/Admixed American; oth = other. Values may underestimate true minimum region-wide percentage, as (i) the least rare such variant (ACE2 p.L731F), which most strongly suppresses the AFR estimate, appears mainly in data from over-proportionately sampled west Africa, more so than in data yet sampled from likewise populous peoples elsewhere in Africa and diaspora [17]; and (ii) any pairwise linkage among shortlisted variants would increase the proportion of people inheriting neither variant in such pairs

Full size image

In principle, real populations may indeed differ, if modestly, in functionally relevant patterns of variation in a gene (or even genome-wide), pending their histories. All else equal, for example, big populations tend to accrue and keep more genetic variation (especially if non-harmful) than do small populations [12, 18, 19]; fast-growing populations to accrue more new rare variation in particular [9, 20]; and genome segments under strong directional selection, in a given population’s environment, may (along with flanking segments coupled to them without recombination) tend to lose such rare variation in particular [10].

But as human genes thus vary in summary patterns of variation, via potentially population-distinctive histories, meaningfully comparing such patterns requires sampling well, and assessing not just which variants appear at all in data sampled from a population, but the summary distributions of their estimated frequencies (respective proportions of sampled gene copies harboring them).

To that end, frequency-sensitive summary metricsFootnote 7 show less variation in human ACE2, both within and between most human populations,Footnote 8 than for most other X-borneFootnote 9 or autosomal [21, 22, 28, 29] human genes, limiting the extent to which populations’ distinctive histories may yield disparate patterns of variation. By comparison, such well grounded summary metrics show more overall variation in human TMPRSS2 [21]—much of it shared across populations, in varied patterns that reflect the cross-regional spread of variants old (and generally non-harmful) enough to have become common.

Importantly, even beyond the two genes’ contrasting patterns of variation, pandemic-long cohort outcomes have not shown variation in either ACE2 or TMPRSS2 to shape personal COVID risks nearly as significantly as variation elsewhere in our genomes—including the most strongly and significantly risk-shaping locus, on the short arm of chromosome 3; the ABO blood group locus on chromosome 9; and other autosomal loci [1,2,3,4,5, 17, 30]. Some non-protein-altering variants in ACE2 and TMPRSS2 have met multiple-test-stringent significance criteria for association with risks of SARS-CoV2 infection (an ACE2 regulatory variant cluster) or severity (TMPRSS2 intronic variant), but their significance falls short of that evident for other loci. And among variants shortlisted by Hou et al., only one (the relatively common TMPRSS2 p.V160M) has shown even suggestive (not multiple-test-stringent) evidence for association with any COVID risk [30,31,32,33]—while broader tests, tuned and powered specifically to detect rare variant association per se in clinically characterized population cohorts, have not implicated shortlisted or other rare protein-coding variation in either gene in COVID risks [34, 35].

Moreover, across all COVID-implicated loci, risk-informative variants differ in their population distributions and inferred effects, in many cases in partial counter-balance to one another. The ACE2-regulatory rs190509934C variant, for example, associates significantly with below-average risk of SARS-CoV2 infection (and suggestively with below-average risk of severe COVID), but appears least rare in a studied broad population (south Asian) in which other loci harbor variant loads most strongly associated with above-average risk of infection (and severe disease) [36, 37].

Altogether, such mixes of above- and below-average risk-associated variants in human genomes worldwide leave a broad range of risk-relevant personal variant load within every studied human population [38,39,40,41]. And those loads, in turn, explain < 10% of personal variability in COVID risks evident, so far, in clinically characterized cohorts [39], while other factors, such as age, background health, and immune exposures, show far stronger effects.Footnote 10 As such, now-ample data suggest that COVID incidence and severity likely differ among human populations far less by genetics than by factors such as age structure, past and ongoing immune exposures, comorbidity prevalences, and access to effective health interventions [4, 17, 30, 39, 40, 45,46,47,48,49,50,51,52]—and pointedly, do not support speculation that variation in ACE2 or TMPRSS2 drives net population disparity in genetically attributable (or overall) COVID risks.Footnote 11

Conclusions

Speculating early in the pandemic, on potential COVID-relevance of variation in two of humanity’s many genes, Hou et al. understandably settled for predictive heuristics in lieu of clinical data. But in tallying shortlisted ACE2 and TMPRSS2 variants among populations, they mistook a sample size artifact as if evidence of population differences—and tried to proxy population-representative gene function by tallying shortlisted variants found at all, instead of summing empirically estimated, genotype frequency-scaled effects. Though their shortlist offered well intended, if unvalidated, candidates for early-pandemic study, genome-wide empirical insights have eclipsed its utility—while leaving, unaddressed, their artifactual summary claims. And those claims, in turn, have drawn credulous citation in public discourse, hindering understanding of COVID risks and of human genetic diversity itself [57,58,59].

While pitfalls of methodology and interpretation have long plagued basic and clinical research [60,61,62,63,64,65,66,67], public discourse invoking Hou et al. [6] highlights how platforms to usefully share and discuss such research can also virally spread faulty inferences missed by authors and reviewers, misleading not just researchers and clinicians, but also lay-people who may rely on published science in personal, professional, family, and civic decisions [64]. As such, we hope that correcting basic errors that misled Hou et al., and those citing their work, helps right the record on human genomic variation in ACE2, TMPRSS2, and loci more informative of COVID risks—and, further, encourages critical stringency in the interpretation of population genetic and other data, amid efforts [68, 69] to more promptly and soundly validate research published, and subsequently invoked, in and between societal crises.

Availability of data and materials

Datasets analyzed in this study are available at:

• https://cloud.google.com/life-sciences/docs/resources/public-datasets/gnomad (gnomAD v3.0 data includes 1000Genomes data, as analyzed by Hou et al.)

• https://esp.gs.washington.edu/drupal/dbGaP_Releases

• https://www.covid19hg.org/

Notes

  1. Denoted by Hou et al. as, respectively, “African/African-American” and “Non-Finnish [and implicitly non-Amish, non-Ashkenazi] European” in Hou et al., these classifications group genetically and culturally disparate peoples, across wide modern and ancestral geographic ranges, at different scales, under single umbrella terms.

  2. Compounding their lopsided sampling, Hou et al. also reported adding, to the foregoing totals, duplicate copies of many non-Amish, non-Ashkenazi genotypes in particular (from the 1000 Genomes Project dataset, which their main gnomAD v3.0 data already included), further confounding any frequency estimation of individual variants.

  3. While cumulative discovery of selectively neutral variants, in a steady-sized randomly mating population, tends to grow logarithmically with sample size [7], Hou et al. instead tallied selection-liable (protein-altering) variants, non-cumulatively (i.e., via independent single samplings) from mutually diverged, variably sized, and potentially variably growing populations. The foregoing factors more conservatively suit linear than logarithmic (or other presumptive higher-order) regression [8,9,10].

  4. Further, each ACE2 variant tally, and all but two TMPRSS2 variant tallies (those for the highly recently admixed “Admixed Latino” and “Population Not Assigned” samples), accords the 95%-confident expectation for a random sample of its respective size, without replacement, from total genotypes studied for that gene, given total shortlisted variant incidences.

  5. Confidence bound presumes accurate estimate of variant frequency in non-Finnish, non-Amish, non-Ashkenazi Europeans, and large randomly mating populations such that sampling probabilities approximate binomial expectation.

  6. Estimates (products of binomial probabilities) presume that variants assort randomly, independently, at sampled population-specific frequencies, in half-XX/half-XY populaces.

  7. For example, copy-pairwise per-site mismatch probability (nucleotide diversity), Tajima’s D, and Wright’s FST.

  8. Save for in mutually neighboring Baka, Bakola, and Bedzan peoples, among whom three protein-coding variants rare or absent in other central African and broader populations have become moderately common (perhaps via shared linkage to regulatory variants favored by local selection over many past generations) [21]. A smaller survey [22] that reported some metrics in European-Americans suggested excess intermediate-frequency variation; but those metrics tend to spuriously infer such skew in small samples [23]. By contrast, in the same study [21], metrics robust to sample size [23] affirmed less overall variation in ACE2 than in other X-linked genes.

  9. Overall, X-specific genes such as ACE2 tend, theoretically and empirically, to vary less among people overall or within a particular population, but more (relative to such scant overall variation) between randomly chosen members of different populations, than do autosomal genes [24,25,26,27].

  10. Even for severe COVID (Hou et al.’s focus), sensitive liability estimates [42] at conservatively estimated population prevalences [43] were < 10% [30]. One non-genetic (pedigree-based) survey suggested possibly greater heritability, still surpassed by inferred effects of residence environment alone, but in lacking genetic data, could not control for close-kin or/and family size biases in exposures within, between, and beyond residences (e.g., in room-sharing, social/care visitation, and/or similarity of work), or for family-wise differences in age or non-residence environmental exposures [44]. And on the specific question of population (versus personal or family) disparity in inborn severe COVID risk, early speculation from sparsely typed small samples [38] has not borne out in well powered cohort data [39, 45].

  11. Specifically, empirically estimated variance in population-mean genomically attributable risk (the summed product of significantly risk-associated variants’ sampling-inferred genotype frequencies and empirically estimated additive risk contributions [36]), for the focal outcome of Hou et al. (severe COVID) [6], in the same populations they studied, is slightly greater (~ 0.02089) when omitting ACE2 variation than when including it (~ 0.02004) or when also including TMPRSS2 variation (0.01822) [37]. These estimates warrant much caution, as variant-attributable effects themselves vary among populations [30] (plausibly via population-stratified genome-wide variation, local haplotype structure, environmental modifiers [53,54,55,56], and other factors), but suggest that, all else equal, ACE2 and (if relevant) TMPRSS2 variation unlikely exacerbates any net disparity in such additively genomically attributable COVID risk among studied human populations.

Abbreviations

ACE2 :

Angiotensin converting enzyme 2 [gene]

AMI:

Amish

AMR:

[Latina/o] Admixed American

ASH:

Ashkenazi

COVID:

Coronavirus disease

DNA:

Deoxyribonucleic acid

EAS:

East Asian

EUR:

[Non-Amish, non-Ashkenazi, non-Finnish] European

FIN:

Finnish

oth:

Other [population(s)]

SARS-CoV2:

Severe acute respiratory syndrome coronavirus 2

SAS:

South Asian

TMPRSS2 :

Transmembrane serine protease 2 [gene]

References

  1. Severe Covid-19 GWAS Group, Ellinghaus D, Degenhardt F, Bujanda L, Buti M, Albillos A, et al. Genomewide association study of severe COVID-19 with respiratory failure. N Engl J Med. 2020;383:1522–34.

    Article  Google Scholar 

  2. Zhao J, Yang Y, Huang H, Li D, Gu D, Lu X, et al. Relationship between the ABO blood group and the COVID-19 susceptibility. Clin Infect Dis. 2021;73(2):328–31.

    Article  PubMed  CAS  Google Scholar 

  3. Pairo-Castineira E, Clohisey S, Klaric L, Bretherick AD, Rawlik K, Pasko D, et al. Genetic mechanisms of critical illness in COVID-19. Nature. 2021;591:92–8.

    Article  PubMed  Google Scholar 

  4. COVID-19 Host Genetics Initiative. Mapping the human genetic architecture of COVID-19. Nature. 2021;600:472–7.

    Article  Google Scholar 

  5. Kousathanas A, Pairo-Castineira E, Rawlik K, Stuckey A, Odhams CA, Walker S, et al. Whole-genome sequencing reveals host factors underlying critical COVID-19. Nature. 2022;607:97–103.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  6. Hou Y, Zhao J, Martin W, Kallianpur A, Chung MK, Jehi L, et al. New insights into genetic susceptibility of COVID-19: an ACE2 and TMPRSS2 polymorphism analysis. BMC Med. 2020;18:216.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  7. Ewens WJ. The sampling theory of selectively neutral alleles. Theor Popul Biol. 1972;3:87–112.

    Article  PubMed  CAS  Google Scholar 

  8. Desai MM, Nicolaisen LE, Walczak AM, Plotkin JB. The structure of allelic diversity in the presence of purifying selection. Theor Popul Biol. 2012;81:144–57.

    Article  PubMed  Google Scholar 

  9. Slatkin M, Hudson RR. Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics. 1991;129:555–62.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  10. Przeworski M, Coop G, Wall JD. The signature of positive selection on standing genetic variation. Evolution. 2005;59:2312–23.

    Article  PubMed  Google Scholar 

  11. Kiezun A, Garimella K, Do R, Stitziel NO, Neale BM, McLaren PJ, et al. Exome sequencing and the genetic basis of complex traits. Nat Genet. 2012;44:623–30.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  12. Fisher RA. The genetical theory of natural selection. 1930.

  13. Mathieson I, Reich D. Differences in the rare variant spectrum among human populations. PLoS Genet. 2017;13:e1006581.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Niroula A, Vihinen M. How good are pathogenicity predictors in detecting benign variants? PLoS Comput Biol. 2019;15:e1006481.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  15. Gunning AC, Fryer V, Fasham J, Crosby AH, Ellard S, Baple E, et al. Assessing performance of pathogenicity predictors using clinically-relevant variant datasets. bioRxiv. 2020;:2020.02.06.937169.

  16. de Garcia FAO, de Andrade ES, Palmero EI. Insights on variant analysis in silico tools for pathogenicity prediction. Front Genet. 2022;13:1010327.

    Article  PubMed  PubMed Central  Google Scholar 

  17. COVID-19 Host Genetics Initiative. A first update on mapping the human genetic architecture of COVID-19. Nature. 2022;608:E1–10.

    Article  Google Scholar 

  18. Simons YB, Turchin MC, Pritchard JK, Sella G. The deleterious mutation load is insensitive to recent population history. Nat Genet. 2014;46:220–4.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  19. Henn BM, Botigué LR, Bustamante CD, Clark AG, Gravel S. Estimating the mutation load in human genomes. Nat Rev Genet. 2015;16:333–43.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  20. Nelson MR, Wegmann D, Ehm MG, Kessner D, St Jean P, Verzilli C, et al. An abundance of rare functional variants in 202 drug target genes sequenced in 14,002 people. Science. 2012;337:100–4.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  21. Zhang C, Verma A, Feng Y, Dos Reis Melo MC, McQuillan M, Hansen M, et al. Impact of natural selection on global patterns of genetic variation, and association with clinical phenotypes, at genes involved in SARS-CoV-2 infection. Res Sq. 2021. https://doi.org/10.21203/rs.3.rs-673011/v1.

  22. Akey JM, Eberle MA, Rieder MJ, Carlson CS, Shriver MD, Nickerson DA, et al. Population history and natural selection shape patterns of genetic variation in 132 genes. PLoS Biol. 2004;2:e286.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Subramanian S. The effects of sample size on population genomic analyses—implications for the tests of neutrality. BMC Genomics. 2016;17:123.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Vicoso B, Charlesworth B. Evolution on the X chromosome: unusual patterns and processes. Nat Rev Genet. 2006;7:645–53.

    Article  PubMed  CAS  Google Scholar 

  25. Akey JM, Zhang G, Zhang K, Jin L, Shriver MD. Interrogating a high-density SNP map for signatures of natural selection. Genome Res. 2002;12:1805–14.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  26. Lambert CA, Connelly CF, Madeoy J, Qiu R, Olson MV, Akey JM. Highly punctuated patterns of population structure on the X chromosome and implications for African evolutionary history. Am J Hum Genet. 2010;86:34–44.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  27. Elhaik E. Empirical distributions of F(ST) from large-scale human polymorphism data. PLoS One. 2012;7:e49837.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  28. Suryamohan K, Diwanji D, Stawiski EW, Gupta R, Miersch S, Liu J, et al. Human ACE2 receptor polymorphisms and altered susceptibility to SARS-CoV-2. Commun Biol. 2021;4:475.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  29. Keinan A, Reich D. Human population differentiation is strongly correlated with local recombination rate. PLoS Genet. 2010;6:e1000886.

    Article  PubMed  PubMed Central  Google Scholar 

  30. COVID-19 Host Genetics Initiative. A second update on mapping the human genetic architecture of COVID-19. Nature. 2023;621:E7–26.

    Article  Google Scholar 

  31. Li W, Zhang C, Sui J, Kuhn JH, Moore MJ, Luo S, et al. Receptor and viral determinants of SARS-coronavirus adaptation to human ACE2. EMBO J. 2005;24(8):1634–43. https://doi.org/10.1038/sj.emboj.7600640.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  32. David A, Parkinson N, Peacock TP, Pairo-Castineira E, Khanna T, Cobat A, et al. A common TMPRSS2 variant has a protective effect against severe COVID-19. Curr Res Transl Med. 2022;70:103333.

    PubMed  CAS  Google Scholar 

  33. Yaghoobi A, Lord JS, Rezaiezadeh JS, Yekaninejad MS, Amini M, Izadi P. TMPRSS2 polymorphism (rs12329760) and the severity of the COVID-19 in Iranian population. PLoS One. 2023;18:e0281750.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  34. Kosmicki JA, Horowitz JE, Banerjee N, Lanche R, Marcketta A, Maxwell E, et al. Pan-ancestry exome-wide association analyses of COVID-19 outcomes in 586,157 individuals. Am J Hum Genet. 2021;108:1350–5.

    Article  PubMed  CAS  Google Scholar 

  35. Butler-Laporte G, Povysil G, Kosmicki JA, Cirulli ET, Drivas T, Furini S, et al. Exome-wide association study to identify rare variants influencing COVID-19 outcomes: results from the Host Genetics Initiative. PLoS Genet. 2022;18:e1010367.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  36. Covid-19 HGI Browser. https://app.covid19hg.org/variants. Accessed 4 Aug 2023.

  37. gnomAD. https://gnomad.broadinstitute.org/. Accessed 23 Oct 2023.

  38. Marçalo R, Neto S, Pinheiro M, Rodrigues AJ, Sousa N, Santos MAS, et al. Evaluation of the genetic risk for COVID-19 outcomes in COPD and differences among worldwide populations. PLoS One. 2022;17:e0264009.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Farooqi R, Kooner JS, Zhang W. Associations between polygenic risk score and COVID-19 susceptibility and severity across ethnic groups: UK Biobank analysis. BMC Med Genomics. 2023;16:150.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  40. Zhu D, Zhao R, Yuan H, Xie Y, Jiang Y, Xu K, et al. Host genetic factors, comorbidities and the risk of severe COVID-19. J Epidemiol Glob Health. 2023;13:279–91.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Córdova-Palomera A, Siffel C, DeBoever C, Wong E, Diogo D, Szalma S. Assessing the potential of polygenic scores to strengthen medical risk prediction models of COVID-19. PLoS One. 2023;18:e0285991.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Ojavee SE, Kutalik Z, Robinson MR. Liability-scale heritability estimation for biobank studies of low-prevalence disease. Am J Hum Genet. 2022;109:2009–17.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  43. Anesi GL. COVID-19: epidemiology, clinical features, and prognosis of the critically ill adult. In: Manaker S, Finlay G, editors. UpToDate. Waltham: UpToDate; 2024.

    Google Scholar 

  44. Brown KL, Ramlall V, Zietz M, Gisladottir U, Tatonetti NP. Estimating the heritability of SARS-CoV-2 susceptibility and COVID-19 severity. Nat Commun. 2024;15:367.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  45. Crossfield SSR, Chaddock NJM, Iles MM, Pujades-Rodriguez M, Morgan AW. Interplay between demographic, clinical and polygenic risk factors for severe COVID-19. Int J Epidemiol. 2022;51:1384–95.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Romero Starke K, Reissig D, Petereit-Haack G, Schmauder S, Nienhaus A, Seidler A. The isolated effect of age on the risk of COVID-19 severe outcomes: a systematic review with meta-analysis. BMJ Glob Health. 2021;6:e006434.

    Article  PubMed  Google Scholar 

  47. Derqui N, Koycheva A, Zhou J, Pillay TD, Crone MA, Hakki S, et al. Risk factors and vectors for SARS-CoV-2 household transmission: a prospective, longitudinal cohort study. Lancet Microbe. 2023;4:e397–408.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  48. Antos A, Kwong ML, Balmorez T, Villanueva A, Murakami S. Unusually high risks of COVID-19 mortality with age-related comorbidities: an adjusted meta-analysis method to improve the risk assessment of mortality using the comorbid mortality data. Infect Dis Rep. 2021;13:700–11.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Gazit S, Shlezinger R, Perez G, Lotan R, Peretz A, Ben-Tov A, et al. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) naturally acquired immunity versus vaccine-induced immunity, reinfections versus breakthrough infections: a retrospective cohort study. Clin Infect Dis. 2022;75:e545–51.

    Article  PubMed  Google Scholar 

  50. Madhi SA, Kwatra G, Myers JE, Jassat W, Dhar N, Mukendi CK, et al. Population immunity and COVID-19 severity with omicron variant in South Africa. N Engl J Med. 2022;386:1314–26.

    Article  PubMed  CAS  Google Scholar 

  51. Mongin D, Bürgisser N, Laurie G, Schimmel G, Vu D-L, Cullati S, et al. Effect of SARS-CoV-2 prior infection and mRNA vaccination on contagiousness and susceptibility to infection. Nat Commun. 2023;14:5452.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  52. Hammond J, Leister-Tebbe H, Gardner A, Abreu P, Bao W, Wisemandle W, et al. Oral nirmatrelvir for high-risk, nonhospitalized adults with COVID-19. N Engl J Med. 2022;386:1397–408.

    Article  PubMed  CAS  Google Scholar 

  53. Martin AR, Gignoux CR, Walters RK, Wojcik GL, Neale BM, Gravel S, et al. Human demographic history impacts genetic risk prediction across diverse populations. Am J Hum Genet. 2017;100:635–49.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  54. Martin AR, Kanai M, Kamatani Y, Okada Y, Neale BM, Daly MJ. Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet. 2019;51:584–91.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  55. Coop G. Reading tea leaves? Polygenic scores and differences in traits among groups. arXiv [q-bio.GN]. 2019.

  56. Polygenic Risk Score Task Force of the International Common Disease Alliance. Responsible use of polygenic risk scores in the clinic: potential benefits, risks and gaps. Nat Med. 2021;27:1876–84.

    Article  CAS  Google Scholar 

  57. Hodgman L, Kennedy RF Jr. denies comments on ‘ethnically targeted’ COVID-19 were anti-Semitic. POLITICO. https://www.politico.com/news/2023/07/15/rfk-jr-covid-19-ethnically-targeted-00106478. Accessed 6 Aug 2023.

  58. Kennedy RF Jr. [...]a 2021 study of the COVID-19 virus [sic] shows that COVID-19 appears to disproportionately affect certain races […] That study is here: https://pubmed.ncbi.nlm.nih.gov/32664879/ Twitter. 2023. https://twitter.com/RobertKennedyJr/status/1680227322509635595. Accessed 4 Aug 2023.

  59. Owens C, Kennedy RF Jr. mentioned an actual scientific study which revealed that Chinese people & Ashkenazi Jews were the most immune to COVID-19 [...]. https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-020-01673-z … Twitter. 2023. https://twitter.com/RealCandaceO/status/1680371443446722561. Accessed 6 Aug 2023.

  60. Remschmidt C, Wichmann O, Harder T. Frequency and impact of confounding by indication and healthy vaccinee bias in observational studies assessing influenza vaccine effectiveness: a systematic review. BMC Infect Dis. 2015;15:429.

    Article  PubMed  PubMed Central  Google Scholar 

  61. Border R, O’Rourke S, de Candia T, Goddard ME, Visscher PM, Yengo L, et al. Assortative mating biases marker-based heritability estimators. Nat Commun. 2022;13:660.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  62. Høeg TB, Duriseti R, Prasad V. Potential, “healthy vaccinee bias” in a study of BNT162b2 vaccine against COVID-19. N Engl J Med. 2023;389:284–6.

    Article  PubMed  Google Scholar 

  63. Herrera-Perez D, Haslam A, Crain T, Gill J, Livingston C, Kaestner V, et al. A comprehensive review of randomized clinical trials in three medical journals reveals 396 medical reversals. Elife. 2019;8:e45183.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  64. Høeg TB, Ladhani S, Prasad V. How methodological pitfalls have created widespread misunderstanding about long COVID. BMJ Evid Based Med. 2023. https://doi.org/10.1136/bmjebm-2023-112338.

    Article  PubMed Central  Google Scholar 

  65. Slutsky DJ. Statistical errors in clinical studies. J Wrist Surg. 2013;2:285–7.

    Article  PubMed  PubMed Central  Google Scholar 

  66. Günel Karadeniz P, Uzabacı E, Atış Kuyuk S, Kaskır Kesin F, Can FE, Seçil M, et al. Statistical errors in articles published in radiology journals. Diagn Interv Radiol. 2019;25:102–8.

    Article  PubMed  Google Scholar 

  67. Dowty JG, Win AK, Buchanan DD, Lindor NM, Macrae FA, Clendenning M, et al. Cancer risks for MLH1 and MSH2 mutation carriers. Hum Mutat. 2013;34:490–7.

    Article  PubMed  CAS  Google Scholar 

  68. Simundic A-M, Nikolac N. Statistical errors in manuscripts submitted to Biochemia Medica journal. Biochem Med. 2009;19:294–300.

    Article  Google Scholar 

  69. Waltman L, Kaltenbrunner W, Pinfield S, Woods HB. How to improve scientific peer review: four schools of thought. Learn Publ. 2023;36:334–47.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank Arjun Biddanda, Charleston Chiang, Mark Iles, Ann Morgan, Richard Sever, reviewers, and preprint commenters for helpful insights and suggestions; and Konrad Karczewski for guidance in accessing the gnomAD v3.0 data used by Hou et al.

Funding

NIH NIGMS grant R35 GM149521 (JN).

Author information

Authors and Affiliations

  1. Root Deep Insight, Boston, MA, USA

    Nathaniel M. Pearson

  2. Department of Human Genetics, University of Chicago, Chicago, IL, USA

    John Novembre

Authors
  1. Nathaniel M. Pearson
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. John Novembre
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

NMP analyzed data, and conceived and drafted the manuscript; JN critically refined analysis and manuscript with NMP. Both authors read and approved the final manuscript.

Authors’ Twitter handles

@genomenathan (NMP), @jnovembre (JN).

Corresponding author

Correspondence to Nathaniel M. Pearson.

Ethics declarations

Ethics approval and consent to participate

N/A.

Consent for publication

N/A.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Pearson, N.M., Novembre, J. No evidence that ACE2 or TMPRSS2 drive population disparity in COVID risks. BMC Med 22, 337 (2024). https://doi.org/10.1186/s12916-024-03539-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12916-024-03539-0

Keywords

BMC Medicine

ISSN: 1741-7015

Contact us