A Haunted GWAS: The Missing Historical and Social Context in the “Gay Gene” Study

Authors: Jonathan Galka, Annika Gompers, and Juanis Becerra

In late August, Science published the results of the largest genome-wide association study (GWAS) to date identifying genes associated with same-sex sexual behavior.[i] The study claims to find five SNPs that are significantly associated with ever having had a same-sex partner, a finding that ignited a wide debate among geneticists and gender and sexual minorities. While each of these SNPs individually accounted for less than 1% of variance in the “trait” of homosexual behavior, the authors state that when considered in aggregate, they account for 8 to 25% of variance - indicating that both genetic and environmental factors influence sexual behavior. Media outlets have since touted this study as revealing that there is no single “gay gene” - instead, multiple genes play a complex and partial role in sexual behavior.

We have responded to Ganna et al. and their claims, calling into question the generalizability of the results of this GWAS.[ii] Here we would like to delve deeper into our critique of the study’s disregard for the impact of historical and social context on their data sets and parameters. Ganna et al. use a particular sample of people in two countries born in a specific time period to make sweeping claims about the genetic basis of same-sex behavior. In doing so, they necessarily treat the single nucleotide polymorphisms (SNPs) (genetic variants) that they found as ahistorical biological objects that carry context-independent information about humans’ collective past, present, and future. We argue that treating SNPs as ahistorical objects is an a priori assumption that leads to numerous oversights with both ethical and scientific consequences. In particular, we believe that the impact of HIV/AIDS on the sample population and the changing social, political, and legal landscape are critical factors for understanding the scope of this GWAS of same-sex sexual behavior and render broad generalizations about the conclusiveness of the study questionable.

Description of the data sets

The researchers conducted their analysis using data from the UK Biobank and a cohort of research participants from 23andMe (primarily located in the United States) to draw their preliminary conclusions. They then used three smaller samples to verify the replicability of their results. Within the available data, the researchers used two parameters to parse their data: they included only data from participants of European descent and within that subsample they dropped all data corresponding to trans, gender non-conforming, and intersex people.

Information on the age and birth year for each data set is included in the table below.

Table 1. Demographic characteristics of samples used in GWAS of same-sex sexuality study

We understand that it is impossible to have a perfectly sampled population. Narrowing the sampled population using ethnic or geodemographic parameters and a cis-gender binary sex model is not unusual within genetic studies. That being said, scientists do have standards and metrics that they use to determine if the data sufficiently represents a population. They also have the ability to narrow or broaden the scope of their claims depending on the generalizability of their sample population. We argue that the methods and standards used by Ganna et al. fail to account for crucially important historical contexts that significantly impacted “non-heterosexual” behavior. We further argue that the researchers should have considered or at the very least recognized these socio-historical events as part of their practice of scientific rigor, especially given that the five SNPs barely met the threshold for statistical significance within scientific standards (p<0.05).

HIV/AIDS Effects on the Sample Population

A glaring contextual omission in Ganna et al.’s study is the historical impact of HIV/AIDS on their sample population. It is a challenge to estimate just how many LGBTQ+ people died during the AIDS epidemic spanning 1987-1997, but estimates of AIDS deaths on a population of people engaging in same-sex sexual behavior can be inferred. By the time the worst of the AIDS epidemic was coming to a close in 1995, one in nine gay men had received an AIDS diagnosis and upwards of 10% of self-identified gay men aged 25-44 had died.[vi] The effect was concentrated in urban centers. At the height of the AIDS epidemic, AIDS-related illness was the leading cause of death among men in general in nearly half of the large cities in the United States.[vii] It has been suggested that it is likely that every gay person alive today who lived through the epidemic between the ages of 25-44 has been touched by the AIDS epidemic in one form or another.[viii] Data from the UK suggests a similar picture, albeit on a smaller scale.[ix] These data suggest a “decimation” of a cohort of gay people born between 1951 and 1970.[x]

The decimation of a cohort born between 1951 and 1970, which by age constitutes the majority of Ganna et al.’s UK BioBank sample and also constitutes a substantial portion of the 23andMe sample, would mean profound consequences for the inferences Ganna et al. attempt to draw between age and incidence of same-sex sexual behavior. It is well known that trends in HIV transmission and AIDS death have changed over time, as the epidemic has shifted from an urban, wealthy, whiter context to one that predominantly affects a more rural, socioeconomically disadvantaged, racial minority population.[xi] Put simply, members of the sample population have experienced the HIV epidemic differently across time and space. Such population stratification - systematic differences in genotypes between subgroups in a population - is often a problem for GWAS, and it is unclear how effectively Ganna et al. accounted for this potential confounding.[xii]

“The central problem is the study’s disregard for the impact of historical and social context on their data sets and parameters.”

That younger individuals reported same-sex sexual behavior at a higher rate than their older counterparts is not simply an artifact of age, but rather it is wrapped up in a recent history of HIV/AIDS wherein a generation of individuals engaging in same-sex sexual behavior variously died or were affected by an epidemic that destroyed their communities.[xiii]

Social, Political, and Legal Landscape

The 1987-1997 AIDS epidemic was but one social trauma nested within a wider sociohistorical continuum experienced by those engaged in same-sex sexual behavior in the 20th and 21st centuries. After World War II, intense post-war homophobic sentiment generated a counter-response in the creation of a number of radical gay liberationist groups in the US throughout the 1960s and 70s. These tensions came to a head in the landmark Stonewall riots of 1969, wherein LGBTQ+ people physically protested police treatment in the streets of Manhattan. The riots were heard across the Atlantic, and as a direct result led to the creation of the UK Gay Liberation Front in 1972. The group would later disband and lead to the Manchester-based Campaign for Homosexual Equality in 1973, which would lead the fight for legal reform.

By the end of the 1970s, LGBTQ+ people and communities were increasingly visible in major cities in the US and the UK.[xiv] Then, the US saw its first documented case of AIDS in April of 1980, retroactively identified as such by the CDC in 1981; the UK saw its first in 1981. For gay men, 72% of AIDS deaths occurred in men aged 25-44 at the beginning of the epidemic.[xv] Throughout the ensuing 1987-1997 epidemic, LGBTQ+ people who grew into adulthood during a period of unprecedented movements for LGBTQ+ liberation began to suffer the combined effects of reinvigorated homophobic rhetoric, illness, and death. Such a disruption in normative expectations among LGBTQ+ people for friendship, family, lifestyle, and lifespan have intense, lasting effects on those who survived.[xvi]

On top of this, the political and legal landscape changed dramatically for LGBTQ+ populations in the US and UK in the late 20th century. Male homosexuality was first decriminalized in the UK in 1967, overturning centuries of legal discrimination.[xvii] The American Psychological Association removed homosexuality from the Diagnostic and Statistical Manual of Mental Disorders (DSM) in 1973,[xviii] and the World Health Organization removed homosexuality from the International Classification of Diseases in 1990.[xix] The United States Supreme Court overturned laws criminalizing sodomy in 2003.[xx] Same-sex marriage was legalized in the UK in 2013 and the US in 2015.

A shifting political landscape significantly impacts individuals’ likelihood of reporting same-sex sexual behavior.[xxi] Individuals who engage in same-sex behavior, particularly of the age groups most heavily sampled by Ganna et al. from both the UK BioBank and 23andMe datasets, have experienced a tumultuous social, legal, and cultural landscape in their lifetimes exacerbated by ongoing stigma, the HIV/AIDS crisis, and multi-level legal battles for equal rights. Fear of discrimination or criminalization are established barriers to reporting LGBTQ+ sexual orientation.[xxii] Although the datasets used by Ganna et al. collected their data in the 21st century, under somewhat improved social conditions for sexual minorities, studies have shown that more accepting legal and political landscapes for same-sex behavior do not bring an end to anxieties about lived prejudice and unequal status.[xxiii] Ganna et al. failed to account for the fact that their sample demographic, individuals who engage in same-sex sexual behavior, is literally haunted by the ghosts of an epidemic that devastated communities across the US and UK. That epidemic, along with the milieu of social, legal, and political factors that have shaped the entwined experiences of LGBTQ+ oppression and liberation over the 20th and 21st centuries, has generated challenges to estimating uniform conditions for self-reporting same-sex sexual behavior across space and time. The GWAS carried out by Ganna et al. is wrongly predicated on assumptions about the categorical stability of “heterosexual” and “non-heterosexual” over time.

Conclusion

As GWAS studies proliferate, we urge researchers to take into consideration the historical, political, and social contexts across the lifetime of their participants. Ganna et al. do recognize that their sample is biased by age (consisting of older individuals) and that reported same-sex behavior increases by birth year. But their only stated consequence of this limitation is that since their sample is not representative of the general population, their findings may not apply to individuals of other ages.[xxiv]

We argue that Ganna et al.’s failure to account for the historical context of their samples poses a serious limitation to their findings. The importance of considering socio-historical factors in genetic studies is not just a theoretical point. Taking these factors into consideration could materially change the results of such studies, and indeed shed more light on the lived experiences of “non-heterosexuals” than blindly feeding data into an algorithm.

Further, a geneticist interested in performing a GWAS might argue that the scope and aim of a GWAS is to investigate the genetic determinants of a phenotype stripped of its environmental context. While we do not suggest that that GWAS studies are completely without merit, this is precisely why they are perhaps categorically insufficient in answering questions that are historically and socially contingent.

Authorship Statement:

Galka and Becerra generated the idea for the blog post topic. Galka, Gompers, and Becerra drew up the initial draft. Galka and Gompers integrated lab member comments and edits during the revision process. All authors provided substantive contributions to the ideas expressed in this blog post and participated in the preparation of the post.

Endnotes:

[i] Ganna, A., Verweij, K. J. H., Nivard, M. G., Maier, R., Wedow, R., Busch, A. S., … Zietsch, B. P. (2019). Large-scale GWAS reveals insights into the genetic architecture of same-sex sexual behavior. Science, 365(6456). doi: 10.1126/science.aat7693

[ii] Richardson, S. et al. Genome studies must account for history. Science, 366(6472) doi: 10.1126/science.aaz6594

[iii] Question marks in this column indicate that Ganna et al. and the original data sources do not disclose the date of collection.

[iv] Question marks in this column indicate that we calculated these birth years using the date of collection and information on participant ages for each sample.

[v] Sanders, A. R., Beecham, G. W., Guo, S., Dawood, K., Rieger, G., Badner, J. A., … Martin, E. R. (2017). Genome-Wide Association Study of Male Sexual Orientation. Scientific Reports, 7(1). doi: 10.1038/s41598-017-15736-4

[vi] Gagnon, J. H., Nardi, P. M., & Levine, M. P. (1997). Introduction. In In changing times: Gay men and lesbians encounter HIV/AIDS (pp. 1–22). Chicago, IL: The University of Chicago Press.

[vii] Ibid.

[viii] Herdt, G. H., & De Vries, B. (2004). From a far place: Social and cultural considerations about HIV among midlife and older gay men. In Gay and lesbian aging: Research and future directions (pp. 73–93). New York, NY: Springer.

[ix] HIV & AIDS Diagnoses and Deaths by Year in the UK. (2019).

[x] Rosenfeld, D., Bartlam, B., & Smith, R. D. (2012). Out of the Closet and Into the Trenches: Gay Male Baby Boomers, Aging, and HIV/AIDS. The Gerontologist, 52(2), 255–264. doi: 10.1093/geront/gnr138

[xi] For information on this demographic transition of the epidemic, see CDC Fact Sheet: Today’s HIV/AIDS Epidemic. (2016)., and HIV in the United States by Geography. (2017) For a historical perspective on shifting trends in HIV infection and AIDS death among Ganna et al.’s sample population, see also: Rosenfeld et al. (2012).

[xii] Reilly, S. (2019, August 29). Retrieved from https://www.broadinstitute.org/blog/opinion-all-science-what-did-we-learn

[xiii] Rosenfeld, D., Bartlam, B., & Smith, R. D. (2012). Out of the Closet and Into the Trenches: Gay Male Baby Boomers, Aging, and HIV/AIDS. The Gerontologist, 52(2), 255–264. doi: 10.1093/geront/gnr138

[xiv] Ibid. See also: Weston, K. (1991). Families we choose: Lesbians, gays, kinship. New York, NY: Columbia University Press. For this citation and the preceding citations regarding the United States, see: Chauncey, G. (1995). Gay New York: the making of the gay male world, 1890-1940. London: Flamingo.

[xv] Rosenfeld, D., Bartlam, B., & Smith, R. D. (2012). Out of the Closet and Into the Trenches: Gay Male Baby Boomers, Aging, and HIV/AIDS. The Gerontologist, 52(2), 255–264. doi: 10.1093/geront/gnr138

[xvi] Ibid. See also: Thompson, E. H. (2004). Expressions Of Manhood: Reconciling Sexualities, Masculinities, And Aging. The Gerontologist, 44(5), 714–719. doi: 10.1093/geront/44.5.714

[xvii] Sexual Offences Act 1967

[xviii] Lyons, R. D. (1973, December 16). Psychiatrists, in a Shift, Declare Homosexuality No Mental Illness. The New York Times, p. 1.

[xix] Cochran, S. D., Drescher, J., Kismödi, E., Giami, A., García-Moreno, C., Atalla, E., … Reed, G. M. (2014). Proposed declassification of disease categories related to sexual orientation in the International Statistical Classification of Diseases and Related Health Problems (ICD-11). Bulletin of the World Health Organization, 92(9), 672–679. doi: 10.2471/blt.14.135541

[xx] Lawrence v. Texas, 539 U.S. 558 (2003).

[xxi] Charlton, B. M., Corliss, H. L., Spiegelman, D., Williams, K., & Austin, S. B. (2016). Changes in Reported Sexual Orientation Following US States Recognition of Same-Sex Couples. American Journal of Public Health, 106(12), 2202–2204. doi: 10.2105/ajph.2016.303449

[xxii] Brooks, H., Llewellyn, C. D., Nadarzynski, T., Pelloso, F. C., Guilherme, F. D. S., Pollard, A., & Jones, C. J. (2018). Sexual orientation disclosure in health care: a systematic review. British Journal of General Practice, 68(668). doi: 10.3399/bjgp18x694841

[xxiii] Fingerhut, A. W., Riggle, E. D. B., & Rostosky, S. S. (2011). Same-Sex Marriage: The Social and Psychological Implications of Policy and Debates. Journal of Social Issues, 67(2), 225–241. doi: 10.1111/j.1540-4560.2011.01695.x

[xxiv] Genetics of Sexual Behavior. (2019). Retrieved from https://geneticsexbehavior.info/what-we-found/.