The biological bases of gender identity

Gender identity — the inner sense of belonging to a gender — has been the subject of scientific inquiry for over half a century. Biological research does not aim to “explain” or justify the experience of transgender people, but rather to understand the mechanisms that contribute to the development of gender identity in every human being. This article presents the current state of knowledge in genetics, endocrinology, neuroscience, and epigenetics, highlighting both the most robust findings and the methodological limitations that remain.

Introduction

For a long time, it was believed that gender identity was determined exclusively by upbringing and the social environment. Starting in the second half of the twentieth century, a growing body of evidence led the scientific community to recognize that biological factors — genetic, hormonal, and neuroanatomical — play a significant role in its development [1]. Studying these bases is not equivalent to seeking a “cause” of being transgender, but rather to understanding the complexity of a human trait that, like many others, emerges from the interaction between biology and environment.

Scientific interest in the biological bases of gender identity is situated within a broader framework: research on sexual development and brain differentiation. The data collected come from diverse fields — from molecular genetics to twin studies, from neuroimaging to endocrinology — and converge toward a multifactorial model in which no single element is sufficient to explain an individual’s gender identity [1].

Genes and chromosomes: beyond XX and XY

The textbook model associating biological sex with only two chromosomal configurations — 46,XX for females and 46,XY for males — is a useful but incomplete simplification. In nature, numerous variations of the sex chromosomes exist, demonstrating that sexual development is a more complex process than a neat dichotomy.

Chromosomal variations

Differences of sex development (DSD), also commonly referred to as intersex conditions, encompass a wide range of conditions. Klinefelter syndrome (47,XXY) affects approximately 1 in 600 individuals assigned male at birth and involves a configuration with an extra X chromosome. Turner syndrome (45,X), in which one of the two sex chromosomes is missing, affects approximately 1 in 2,500 individuals assigned female at birth. There are also mosaicisms (45,X/46,XY), chimeras (46,XX/46,XY), and other less common combinations.

According to the most widely cited estimates in the scientific literature (Fausto-Sterling, 2000), approximately 1.7% of the population is born with characteristics that do not fit conventional definitions of “male” or “female,” although many of these variations are not visible at birth and are discovered only later. This estimate is, however, debated: Sax (2002) proposed a more conservative figure of 0.018% limited to cases with ambiguous genitalia at birth, while intermediate estimates fall around 0.5-1%. The range depends largely on which conditions are included in the definition of “intersex.”

Implications for gender identity

DSD conditions do not directly concern gender identity in the majority of cases, but their existence demonstrates that biological sex is a continuum rather than a binary variable. Complete androgen insensitivity syndrome (CAIS), for example, involves people with a 46,XY karyotype whose tissues do not respond to testosterone: these individuals develop female phenotypic characteristics and in the vast majority of cases identify as women, suggesting that chromosomes alone do not determine gender identity [1].

Prenatal hormones and brain differentiation

One of the most studied hypotheses in this field is the prenatal hormone exposure theory. According to this model, gender identity is influenced by the hormonal environment to which the fetus is exposed during critical windows of brain development, in a manner partially independent of genital differentiation [1][9].

The critical window

During gestation, the fetal brain goes through sensitive periods in which steroid hormones — particularly testosterone and its metabolites — influence the formation of neural circuits related to behavior and, according to some researchers, gender identity. Genital differentiation occurs in the first trimester, while brain differentiation continues in the second and third trimesters. This temporal gap could explain why, in some cases, the direction of brain development diverges from that of genital development [1].

Studies on congenital adrenal hyperplasia

An important source of data comes from studies on individuals with congenital adrenal hyperplasia (CAH), a genetic condition in which the adrenal glands produce elevated levels of androgens during prenatal life. Individuals with CAH assigned female at birth are exposed to higher-than-normal testosterone levels during fetal development.

A study published in the Journal of Clinical Endocrinology and Metabolism (Dessens et al., 2005) found that 5.2% of CAH patients raised as female developed significant problems with their gender identity, a rate far higher than that of the general population. Subsequent research, such as that published in the same journal in 2003, confirmed that girls with CAH show, on average, play interests and behaviors more typically associated with males, while in the vast majority of cases maintaining a female gender identity [5].

These data suggest that prenatal androgens influence gender-related behavior, but are not the sole determining factor of gender identity [5]. The fact that most people with CAH identify with the gender assigned at birth indicates that other factors — genetic, epigenetic, and social — are involved in the process.

Twin studies

Twin studies represent a classic tool of behavioral genetics for estimating the relative contribution of genes and environment to a given trait. Since monozygotic (MZ) twins share 100% of their DNA, while dizygotic (DZ) twins share on average 50%, a higher concordance rate in MZ twins compared to DZ twins suggests a genetic component.

Concordance and heritability estimates

A systematic review of the twin literature, published in the journal Behavior Genetics in 2025, analyzed eight studies yielding heritability estimates ranging from 0.00 to 0.84. Seven of eight studies provided evidence supporting a genetic component, with heritability estimates between 0.10 and 0.81 [2].

A pioneering study by Coolidge et al., published in Behavior Genetics in 2002, analyzed 314 twins (96 MZ pairs and 61 DZ pairs) aged 4 to 17. The model that best described the data included an additive genetic component accounting for 62% of the variance and a nonshared environmental component explaining the remaining 38% [6].

A more recent study, published in Scientific Reports in 2025, estimated the relative risk ratios for transgender concordance: 21.2 for MZ pairs and 8.7 for DZ pairs (based on an estimated prevalence of 1% in the population), suggesting a substantial genetic contribution to gender diversity [14].

Limitations of twin studies

Heritability estimates show considerable variability across studies due to differences in sample sizes, diagnostic criteria, and statistical methods. The nonshared environmental component — which includes all experiential factors unique to each twin — is consistently significant, with estimates between 0.15 and 0.96 [2]. This means that, while a genetic basis exists, the individual environment plays a non-negligible role.

Neuroimaging: differences in brain structure

Neuroimaging techniques, particularly structural magnetic resonance imaging (MRI), have made it possible to investigate whether differences in brain structure exist between cisgender and transgender individuals.

The ENIGMA study

The largest study conducted to date is that of the ENIGMA (Enhancing Neuro Imaging Genetics through Meta-Analysis) working group on transgender individuals, published in the Journal of Sexual Medicine in 2021 [3]. This mega-analysis examined structural MRI data from 803 participants not undergoing hormonal treatment: 214 transgender men, 172 transgender women, 221 cisgender men, and 196 cisgender women.

The results showed that transgender individuals differ significantly from cisgender individuals in (sub)cortical volumes and cortical surface area, but not in cortical thickness. A particularly relevant finding is that the brains of transgender individuals do not simply fall “halfway” between male and female phenotypes, but present a distinct phenotype of their own [3].

Earlier studies

Earlier research, synthesized in a review published in Archives of Sexual Behavior in 2021, had already shown that the brains of transgender women display a complex mixture of regions with male, female, and “demasculinized” characteristics, while those of transgender men show female, male, and “defeminized” regions. A 2016 review, published in PLOS ONE, concluded that the conflicting results across studies made it difficult to identify brain characteristics that consistently differed between cisgender and transgender groups [11].

Methodological limitations

Neuroanatomical research on gender identity has significant limitations. Samples are often small or medium-sized. The distinction between pre-existing brain differences and those potentially shaped by social experience remains an open problem. Furthermore, the human brain cannot be neatly classified as “male” or “female”: most individuals present a mosaic of traits, making problematic any attempt to use neuroimaging as a diagnostic tool [11].

Epigenetics

Epigenetics studies modifications of gene expression that do not involve changes to the DNA sequence. Epigenetic mechanisms — such as DNA methylation and histone modification — can be influenced by environmental factors, including hormones, and are potentially transmissible across cell generations.

DNA methylation and gender identity

A study published in Frontiers in Neuroscience in 2021 conducted a global epigenomic analysis (EWAS, Epigenome-Wide Association Study) comparing the DNA methylation profiles of 16 transgender and 16 cisgender individuals, all before hormonal treatment. The results showed that the two populations have different global CpG methylation profiles, and that the most significant CpG sites were associated with genes involved in central nervous system development [4].

An additional study, published in Scientific Reports in 2023, found that the CBLL1 gene is hypomethylated in transgender men before hormone therapy and that the degree of methylation correlates with cortical thickness, suggesting a possible link between epigenetic regulation and brain structure [13].

A theoretical model

The current working hypothesis proposes that during fetal development, the genitals and brain may differentiate in different directions, and that epigenetics represents the mechanism through which this occurs: hormones, at critical moments of development, induce epigenetic modifications that silence or activate specific genes differently in different tissues [1][4].

It should be noted that the application of epigenetics to the study of sex and gender is a relatively new field. The samples used are still small, and the results, while promising, require large-scale confirmation.

Why there is no single “gender gene”

Unlike some monogenic conditions — where a single variant of a gene is sufficient to cause a phenotype — gender identity appears to be a complex multifactorial trait, influenced by many genes with small effects that interact with each other and with the environment [1].

Genetic association studies

Research seeking genetic variants associated with gender identity has identified some significant associations, but no “gender gene.” A study published in the Journal of Clinical Endocrinology and Metabolism in 2019 found an association between gender dysphoria and allelic variants in the ERa (estrogen receptor alpha), SRD5A2 (5-alpha reductase type 2), and STS (steroid sulfatase) genes, all involved in sex hormone metabolism [7].

Another study, published in Scientific Reports in 2019, used exome sequencing to identify rare variants in transgender individuals, finding mutations in the RYR3 gene in three unrelated transgender individuals [8]. However, these results suggest at most an oligogenic component — that is, attributable to a limited number of genes — rather than simple genetic determinism.

The multifactorial model

The most widely accepted hypothesis in the current scientific literature is that gender identity is a polygenic trait with a heritable component, in which common and rare variants of numerous genes each contribute a modest effect to the overall risk. To these are added epigenetic, prenatal hormonal, and environmental factors, in an interaction that research is still working to delineate. As emphasized in the review by Polderman and colleagues published in Behavior Genetics in 2018, gender identity “likely reflects a complex interplay of biological, environmental, and cultural factors” [1].

This complexity explains why large-scale GWAS (Genome-Wide Association Studies) have not yet identified loci with genome-wide significant effects: much larger samples than those available so far are needed to achieve the necessary statistical power.

Current scientific consensus

The major international scientific and health institutions have weighed in on the biological nature of gender identity, while acknowledging that understanding of the specific mechanisms is still evolving.

Endocrine Society

The Endocrine Society, in its 2017 clinical practice guidelines and updated position statement, states that “considerable scientific evidence has established a durable biological element underlying gender identity” [9][10]. The society emphasizes that, although the specific biological mechanisms are not yet fully understood, results from diverse disciplines — genetics, endocrinology, neuroanatomy — support the concept that gender identity reflects a complex interplay of biological, environmental, and cultural factors [10].

American Psychological Association

The American Psychological Association (APA), in its Task Force report on gender identity and gender variance, describes the origin of transgender identity as “likely the result of a complex interaction between biological and environmental factors.” In 2024, the APA reiterated that transgender and nonbinary identities represent “normal variations in human gender expression” and took a position against attempts to change people’s gender identity.

World Health Organization

The World Health Organization (WHO) took a significant step with the adoption of the ICD-11 in 2019, reclassifying gender incongruence from the chapter on mental disorders to a new chapter dedicated to “conditions related to sexual health” [12]. This also had implications for transgender rights internationally. This reclassification reflects the scientific consensus that transgender identity is not a mental disorder: in the ICD-11 definition, unlike the DSM-5, neither psychological distress nor dysfunction are necessary requirements for diagnosis [12].

An evolving picture

The current scientific consensus can be summarized in several points: gender identity has a significant, though not exclusive, biological basis [1][10]; it is not the result of voluntary choice or external influences; it is a multifactorial trait in which genes, hormones, epigenetics, and environment interact in ways that are yet to be fully clarified [1]. Research continues to advance, and studies with larger samples and more refined methodologies will in the future be able to provide a more detailed picture of these mechanisms.