Whole exome sequencing analysis of susceptibility loci in transgender individuals

Abstract

Background

Genetic factors contributing to sex-associated dimorphic brain development may also underlie gender identity–related anxiety disorders.

Aim

To establish a high-throughput whole-exome sequencing (WES) and bioinformatics pipeline for identifying rare variants in sex-dimorphic neural pathways and explore their association with gender identity–related anxiety.

Methods

Peripheral genomic DNA was collected from 23 patients (13 Assigned male at birth (AMAB), 10 Assigned female at birth (AFAB)) presenting with gender identity–related anxiety at Shanghai Mental Health Centre between March 2020 and February 2022. WES libraries were prepared and sequenced to an average depth of 100×. Raw reads underwent stringent quality control, alignment, variant calling, and annotation against public databases (gnomAD, ClinVar). Rare (minor allele frequency [MAF] < 1%) high-confidence variants were filtered to focus on exonic, splice-site, and insertion and deletion (indel) events. Candidate genes were subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment and Gene Ontology (GO) analyses to identify overrepresented neural development pathways, with particular emphasis on estrogen receptor–mediated signaling.

Outcome

A total of 479 rare, potentially pathogenic variants across 19 estrogen receptor–mediated neurodevelopmental genes were identified for further validation.

Results

After quality control, 266 265 high-confidence variants were retained; of 217 757 exomic calls, 48 508 (22.3%) were exonic (49.8% nonsynonymous), including 4.1% splice-site, 225 missense, 27 nonsense, 43 frameshift, and various indels. KEGG analysis highlighted significant enrichment in axon guidance signaling, while GO terms pointed to neuronal projection and synaptic assembly. Nineteen genes within the estrogen receptor pathway harbored rare deleterious variants, suggesting disruptions in sex hormone–driven neural differentiation.

Clinical Translation

This WES-based framework enables the identification of novel candidate loci for diagnostic panels and may inform personalized interventions for gender identity–related anxiety.

Strengths and Limitations

This study leveraged high-depth whole-exome sequencing, stringent bioinformatics filtering, and a pathway-focused approach to pinpoint rare variants in sex-dimorphic neurodevelopmental genes; however, its small sample size, lack of functional validation, potential population stratification bias, and cross-sectional design limit causal inference.

Conclusion

Our integrated WES and bioinformatics pipeline uncovers rare variants in estrogen receptor–mediated neurodevelopmental genes, providing new insights into the genetic architecture of sex-dimorphic brain development and its role in gender identity–related anxiety.

Study registration

The article was registered at the ISRCTN Registry https://www.isrctn.com/ISRCTN18336816 (no. 18336816) under an observational study record.

Introduction

Gender incongruence (GI) is defined as a marked and persistent incongruence between an individual’s experienced gender and their assigned sex at birth. This term replaces the previous classification of “gender dysphoria” in the International Classification of Diseases, 10th Revision (ICD-10), and is no longer categorized as a mental disorder.^1-3 Instead, it is included in the chapter on conditions related to sexual health. Importantly, the minority-stress model—which attributes higher rates of depression, anxiety, and suicidal ideation among gender-diverse populations to chronic social stigma, prejudice, and discrimination—provides a critical lens for understanding the psychosocial challenges these individuals face.⁴

Terminology in this field is complex and varies across cultures and disciplines. Terms such as “transgender,” “nonbinary,” “gender nonconforming,” and “gender diverse” are often used to describe individuals whose gender identity or expression differs from societal expectations associated with their assigned sex at birth.^5,6 It is important to distinguish between biological sex, which refers to physical and genetic characteristics, and gender, which encompasses the roles, behaviors, and identities that societies attribute to individuals.

The experiences of transgender and gender-diverse individuals are diverse and multifaceted. Some individuals identify within the traditional binary framework (male or female), while others identify outside of this binary, embracing identities such as nonbinary or gender-fluid. These identities are valid and reflect the rich diversity of human gender experiences.^7,8 Research indicates that transgender and gender-diverse individuals often face significant psychosocial challenges, including higher rates of depression, anxiety, and suicidal ideation compared to the general population. For instance, a study focusing on transgender and gender-diverse youth in Australia reported that 74.6% had been diagnosed with depression and 72.2% with anxiety, 82.4% had experienced suicidal thoughts, and 48.1% had attempted suicide.⁹ Similarly, a study involving hospitalized transgender adolescents in the United States found that 91% had mood disorders, 65% had anxiety disorders, and 52.4% reported suicidal ideation, compared to 39.2% in their cisgender peers.¹⁰ These findings underscore the urgent need for targeted mental health support and interventions for transgender and gender-diverse populations.

Emerging research suggests that genetic, hormonal, and neurological factors may contribute to the development of gender identity, although the precise biological mechanisms remain incompletely understood. Whole-exome sequencing (WES) studies have identified rare variants in genes associated with estrogen receptor signaling pathways, which are implicated in the sexual differentiation of the brain. WES of 30 transgender individuals (13 trans men, 17 trans women) identified 21 rare variants in 19 genes linked to estrogen receptor pathways involved in brain sexual differentiation, suggesting a potential genetic basis for gender identity.¹¹ Additionally, polymorphisms in the estrogen receptor α gene (ESR1) have been associated with female-to-male gender incongruence, suggesting a genetic component to gender identity development.¹² Furthermore, epigenetic mechanisms involving estrogen receptor α (ERα) have been shown to play a role in brain sexual differentiation, indicating a complex interplay of genetic and hormonal factors.¹³ However, these findings are preliminary, and further research is necessary to elucidate the multifaceted biological underpinnings of gender identity.

In this study, we aim to explore potential genetic contributions to gender identity by performing WES on genomic DNA from 23 individuals experiencing gender incongruence. Our goal is to identify rare genetic variants in pathways related to brain sexual differentiation, contributing to a more comprehensive understanding of the biological aspects of gender identity.

Methods

Study design and participants

This study, conducted internationally, aimed to identify rare genetic variants associated with gender incongruence by WES. A total of 23 individuals (13 assigned male at birth and 10 assigned female at birth), all of whom self-identified as transgender, diagnosed with gender incongruence were recruited from the Shanghai Mental Health Center between March 2020 and February 2022. The inclusion criteria encompassed individuals aged 18 years or older who met the diagnostic criteria for gender incongruence as outlined in the ICD-10. The exclusion criteria included the presence of known chromosomal abnormalities or endocrine disorders. All the participants provided written informed consent, and the study was approved by the Ethics Committee of Shanghai Mental Health Center, Shanghai, China.

DNA extraction and quality assessment

Peripheral blood samples were collected from each participant, and genomic DNA was extracted using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol. The concentration and purity of the extracted DNA were assessed using the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States), and integrity was evaluated by agarose gel electrophoresis.

Whole-exome sequencing

Whole-exome libraries were prepared using the SureSelect Human All Exon V6 kit (Agilent Technologies, Santa Clara, CA, United States). Sequencing was performed on the Illumina NovaSeq 6000 platform (Illumina, San Diego, CA, United States) to generate 150 bp paired-end reads, aiming for a minimum mean coverage depth of 100× across the targeted exonic regions.

Bioinformatics analysis

Raw sequencing data were processed through a standardized bioinformatics pipeline. First, FastQC (v0.11.9; Babraham Bioinformatics, Cambridge, United Kingdom) was used to assess the quality of raw reads. High-quality reads were then aligned to the human reference genome (GRCh38/hg38) using BWA-MEM (v0.7.17; Wellcome Trust Sanger Institute, Cambridge, United Kingdom). Variant calling was conducted using GATK HaplotypeCaller (v4.1.9.0; Broad Institute, Cambridge, MA, United States) to identify single nucleotide variants (SNVs) and insertion and deletion (indel) variants. Identified variants were annotated using ANNOVAR (2018Apr16; Wang Genomics Lab, University of Pennsylvania, Philadelphia, PA, United States), with annotation sourced from multiple public variant databases, including ExAC, the 1000 Genomes Project, and ESP6500.

Variant filtering

To identify potentially pathogenic variants, we applied a multistep filtering process. Variants with a minor allele frequency (MAF) of less than 0.01 in ExAC, 1000 Genomes, and ESP6500 databases were retained. Only variants classified as pathogenic or likely pathogenic based on the American College of Medical Genetics and Genomics (ACMG) guidelines were considered. Furthermore, we prioritized variants with predicted high functional impact, including frameshift mutations, nonsense mutations, canonical splice-site changes, and missense variants with a Combined Annotation Dependent Depletion (CADD) score of 20 or greater (Figure 1). Variants found within segmental duplications or other high-homology genomic regions were excluded to reduce the likelihood of false-positive findings.¹⁴

Figure 1.

Workflow for filtering and functionally annotating whole-exome sequencing variants. Starting from 266 265 raw variants detected in 23 samples, three sequential filters were applied: (1) minor allele frequency < 1% in ExAC, 1000 genomes, and ESP6500si v2; (2) ACMG class 3/4 variants (including frameshift, splicing, nonsynonymous SNVs, and CADD score ≥ 20); and (3) exclusion of variants in genomicSuperDups. This reduced the set to 1741 candidate variants. Variants absent from dbSNP147/138 (n = 479) were then categorized by type (345 nonsynonymous SNVs, 10 splicing, and 124 frameshifts), and further prioritized by in silico protein function prediction tools-Sorting Intolerant From Tolerant (SIFT), Polymorphism Phenotyping v2 (PolyPhen-2), Functional Analysis through Hidden Markow Models (FATHMM), Likelihood Ratio Test (LRT)—and by evolutionary conservation scores-Genomic Evolutionary Rate Profiling++ (GERP++) phyloP, phastCons analyses-yielding 27 high-confidence nonsynonymous SNVs. Meanwhile, splicing and frameshift events were tallied (7 splices, 43 deletions, and 28 insertions). Finally, KEGG and GO enrichment on the 1741 filtered variants highlighted pathways related to estrogen signaling and neuronal development.

Functional enrichment analysis

Genes harboring filtered, rare, and potentially deleterious variants were subjected to functional enrichment analysis. Gene Ontology (GO) analysis was performed using the DAVID Bioinformatics Resources (v6.8; National Institute of Allergy and Infectious Diseases, Bethesda, MD, United States) to explore enriched biological processes, molecular functions, and cellular components. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment was analyzed using the KEGG Mapper tool (Kyoto University Bioinformatics Center, Kyoto, Japan) to identify affected biological pathways. Empirical P values were calculated by comparing observed enrichment scores with distributions derived from 1000 randomly selected gene sets, matched for gene length and GC content, to control for potential biases.

Results

This study initially enrolled 23 transgender individuals, comprising 13 transgender males (individuals assigned female at birth who transitioned to male) and 10 transgender females (individuals assigned male at birth who transitioned to female). WES achieved a coverage of 98.36% with an average read depth of 75, revealing 266 265 genetic variations. Among these, SNVs accounted for 217 757, with mutations in the exonic region making up 26.26% and nonsynonymous mutations in the exonic region accounting for 49.76%. Additionally, there were 48 508 indels with mutations in the exonic region representing 4.09%, mutations in the splicing region accounting for 0.28%, insertions in the exonic region leading to frameshift mutations at 13.77%, and deletions causing frameshift mutations at 19.40% (Figures 2 and 3). Further investigation was required to identify 1741 variants following screening.

Figure 2.

Mutation function distribution map. Distribution of target single nucleotide variants (SNVs) and insertions/deletions (indels) across genomic regions and their associated functions. The top-left pie chart shows the genomic distribution of target SNVs, highlighting their prevalence in coding regions. The top-right pie chart details the functional classification of exonic SNVs, emphasizing nonsynonymous variants as the most common. The bottom-left pie chart presents the genomic distribution of target indels, demonstrating their occurrence in coding and noncoding regions. Lastly, the bottom-right pie chart breaks down the functional types of exonic indels, with a focus on frameshift mutations. Each chart employs a distinct color scheme to enhance clarity and is accompanied by a legend for straightforward interpretation.

Figure 3.

Circles and distribution of filtered genes from whole exome sequencing. The outer ring displays the density of single nucleotide polymorphisms (SNPs), with higher-density regions indicated by more intense shading. The middle ring shows transitions and transversions, providing insight into the types of genetic variations observed. The inner ring highlights the total number of insertions and deletions (indels), with insertions and deletions represented by distinct visual patterns. The accompanying legend on the right details the total counts of SNPs and indels, as well as specific transition and transversion types. This comprehensive view aids in understanding the genomic landscape of the filtered genes.

KEGG and GO enrichment analyses were conducted on the identified variants, followed by further filtering to retain 479 variants not present in dbSNP. Subsequently, protein function analysis demonstrated that these variants are related to the development of estrogen and estrogen receptor neurons (Figure 4).

Figure 4.

KEGG and GO pathway enrichment analysis of mutant genes affecting susceptibility, and functional enrichment analysis of genes with rare variants from whole exome sequencing. In the KEGG panel, the “axon guidance” pathway is emphasized due to its critical role in neural development. The BP panel highlights “neuron development” and “cell morphogenesis involved in neuron differentiation,”underscoring their significance in neuron specialization. The CC panel spotlights identifies “neuron projection,” as vital for neuron structure and function. Lastly, the MF panel includes “metallopeptidase activity” and “cytoskeletal protein binding,” which are essential for various biological processes. Each panel is displayed using a bubble plot, where bubble size reflects the number of genes and enrichment significance is represented by gradient shading or position along the significance axis (−log10(P value)). Abbreviations: BP, biological process; CC, cellular component; MF, molecular function.

Nonsense variants

After filtering, the 27 nonsensical variants initially identified by WES still remained (Table 1). Each variant was found in a single subject and was heterozygous. Conservative evolutionary analysis revealed ATP7A as an X chromosome mutation, with all genes situated in conserved regions. Furthermore, mutations in AGRN, KLKB1, ABCA1, KCNJ14, and ARFRP1 were positioned at transcription factor binding sites.

Table 1.

Ambiguous variants called through whole exome sequencing.

Gene.refGene	Chrom	Pos	Ref	Alts	Alt	Het (n)	Hom (n)
AGRN	chr1	979016	G	A	A	1	0
CYP1B1	chr2	38298404	C	T	T	1	0
CNTNAP5	chr2	124979332	A	G	G	1	0
ITPR1	chr3	4730250	C	A	A	1	0
KLKB1#	chr4	187179251	G	T	T	1	0
CHRNA9	chr4	40351422	C	T	T	1	0
SLC6A7	chr5	149582254	G	C	C	1	0
ARSI	chr5	149681654	G	A	A	1	0
OPRM1	chr6	154412447	C	G	G	1	0
WASHC5	chr8	126093990	T	C	C	1	0
CTHRC1	chr8	104387987	G	A	A	1	0
ABCA1	chr9	107568541	C	T	T	1	0
NOTCH1	chr9	139412284	T	C	C	1	0
TSPAN14	chr10	82267123	T	C	C	1	0
EXT2	chr11	44148477	C	G	G	1	0
SLC10A2	chr13	103703646	C	A	A	1	0
POMT2	chr14	77746787	G	A	A	1	0
PRKD1	chr14	30135343	A	G	G	1	0
SLC7A7	chr14	23243690	A	G	G	1	0
ATP10A	chr15	25932897	A	G	G	1	0
GALK2	chr15	49493405	C	A	A	1	0
PHKB	chr16	47545585	T	A	A	1	0
EML2	chr19	46119077	T	A	A	1	0
KCNJ14^	chr19	48965151	G	A	A	1	0
ARFRP1	chr20	62338359	C	G	G	1	0
NOL12	chr22	38086786	C	T	T	1	0
ATP7A&	chrX	77287057	G	A	A	1	0

Note: Gene.refGene: according to the RefSeq Gene database, the name of the gene where the mutation site is located; Chrom: chromosome ID; Pos: location on the chromosome, starting from 1; Ref is the reference base. The ALTS software calls out bases that are inconsistent with the reference base. When there is more than one base, it is separated by a comma and displayed in multiple rows, with the ALT in each row being one of the ALTS bases; the ALT software calls out bases that are inconsistent with the reference base, usually consistent with the ALTS column. When there are multiple ALTS, ALT corresponds to a nonreference base in the ALTS and selects the most reasonable base. Het (n) lists the number of heterozygous mutations and Hom (n) the number of homozygous mutations.

Frameshift variants

Among the code shift variants identified by WES, 43 were filtered out (Table S1). Mutations in AKAP12, ZBTB39, ZNF213, RPRD1A, and BLVRB were located at the transcription factor binding site. Additionally, AKAP12, CNOT2, ZNF626, and VSIG4 exhibited homozygous mutations, with ZNF626 showing the highest number of such mutations. The total number of TBP mutations was the highest, and VSIG4 was associated with an X chromosome mutation.

Further analysis revealed that frameshift insertions were identified through WES and mapped to conserved elements based on the 46 way vertebrate alignment (Table S2). Mutations in FOXO6, MAST2, KIAA1109, H1-6, ANKS1A, RERGL, MBTPS1, ABR, TEKT1, and DCAF15 were found to be located at transcription factor binding sites, with CHST15 showing the highest number of frameshift insertion mutations. Additionally, TNFAIP6, CHST15, CNOT2, and DCAF15 were identified as homozygous mutations, with DCAF15 having the highest number of homozygous mutations.

Splice-region variants

Among the splice variants initially identified by WES, 7 were retained after filtration (Table 2). CFAP94, GRIN2C, and CD24 exhibited pure sum mutations, with CFAP94 showing the highest number of mutations. CD24 represents a Y chromosome mutation. SHANK3 phastConsElements46way received the highest and most conservative rating.

Table 2.

Splicing variants called by whole exome sequencing.

Gene.refGene	Chrom	Pos	Ref	Alts	Alt	Het (n)	Hom (n)
HNRNPR*	chr1	23640196	C	CT	CT	1	0
GEMIN5*	chr5	154287380	T	TACA	TACA	1	0
ZC3HAV1	chr7	138764989	GC	G	G	1	0
HMCN2*	chr9	133305073	GTACGGGGACACCCACCCTCTGGCCACACCGCTGCAGCTGCCCCAGGGGTTACCGGATGCAGGGCCCCAGCCTGCCCTGCCTAGT	G	G	1	0
GLT6D1*	chr9	138530956	TATTTAC	T	T	1	0
CFAP94*	chr12	25261759	T	TAAAAAAAAAAAA, TAAAAAAAAAAAAAAAAAAAAA, TAAAAAAAAAAAAAAAAAAAA	TAAAAAAAAAAAAAAAAAAAA	1	4
GRIN2C*	chr17	72848752	T	TG	TG	0	1
SHANK3*^	chr22	51135989	GTT	G	G	1	0
CD24	chrY	21154 705	C	CCAGGAAAGCCACAA, CCAGGAAAGCCACAATAGCCGTGACG	CCAGGAAAGCCACAA	2	2
CD24	chrY	21154705	C	CCAGGAAAGCCACAA, CCAGGAAAGCCACAATAGCCGTGACG	CCAGGAAAGCCACAATAGCCGTGACG	2	2

Missense variants

The mutation identified through whole exome sequencing in three or more patients involves 11 genes (Table 3). Among these genes, the phastConsElements46way score was highest for CNOT2, indicating a more conservative nature. Furthermore, the DCAF15 mutation was found at a transcription factor binding site when compared to the UCSC tfbsConsSite database, with the highest number of TBP and CHST15 mutations. ZNF626 exhibited the highest number of homozygous mutations, while CD24 was associated with a Y chromosome mutation.

Table 3.

Variants called by whole exome sequencing in three or more than three patients.

Gene.refGene	Chrom	Pos	ExonicFunc.refGene	Ref	Alts	Het (n)	Hom (n)	N
TBP	chr6	170871054	Frameshift deletion	AGCAGCAGCAGCAGCAGCAGCAG	ACAGCAGCAGCAGCAGCAGCAG, ACAGCAGCAG, A, ACAGCAG, ACAGCAGCAGCAG	19	3	22
CHST15	chr10	125780752	Frameshift insertion	CG	C, CGGG	16	6	22
TNFAIP6	chr2	152236045	Frameshift insertion	TA	TAA, T	17	3	20
CNOT2^	chr12	70747693	Frameshift deletion/insertion	TAA	T, TA, TAAA	9	6	15
DCAF15·	chr19	14070706	Frameshift insertion	A	AGGTGGGCCCAGGGCGGGCAG, AGGTGTGCCCAGGGCGGGCAG	6	8	14
ANKLE1	chr19	17397456	Frameshift deletion/insertion	GGTGTGTGTGT	GGTGTGT, GGTGTGTGT, GGTGTGTGTGTGT, GGTGTGTGTGTGTGTGT, G	12	0	12
ZNF626	chr19	20807133	Frameshift deletion	GGCTTTGCCACATTCTTCACATTTGTAGAATTTCTCTCCAGTATGATTCTCTCATGTGTAGTAAGGATTGAGGACTGGTTGAAGGCTTTGCCACATTCTTCACATTTGTAGGGTCTCTCTCCAGTATGAATTTTCTTATGTGTAGTAAGGTTAGAGGAGCACTTAAAA	G	0	12	12
DENND4B	chr1	153907306	Frameshift insertion	T	TGC, TGCTGCTGCTGCTGCTGCTGC, TGCTGCTGCTGC	6	0	6
RBM5	chr3	50155887	Frameshift insertion	TGA	T, TGAGAGA	5	0	5
CFAP94	chr12	25261759	Splicing	T	TAAAAAAAAAAAA, TAAAAAAAAAAAAAAAAAAAAA, TAAAAAAAAAAAAAAAAAAAA	1	4	5
CD24	chrY	21154705	Splicing	C	CCAGGAAAGCCACAA, CCAGGAAAGCCACAATAGCCGTGACG	2	2	4
SERPINA1	chr14	94845829	Frameshift insertion	G	GAC	3	0	3

Candidate gene variations are associated with neuronal development involving estrogen and estrogen receptor signaling.

Comparison of mutations in the estrogen and estrogen receptor signaling pathways identified 19 associated genes, including CYP1B1, CNOT2, AGRN, CHRNA9, OPRM1, CTHRC1, WASHC5, NOTCH1, PRKD1, KCNJ14, ATP7A, CCDC141, IFT88, FOXO6, ANKS1A, ABR, FBXO7, HNRNPR, and SHANK3 (Table 4) Among these, CNOT2 was found to be the most evolutionarily conserved. CNOT2 functions as a negative regulator of the estrogen receptor signaling pathway, whereas ATP7A, an X-linked gene, is implicated in neuronal functions.

Table 4.

Variants of candidate genes related to estrogen and estrogen receptor and neuron development.

Relation	ExonicFunc.refGene	Gene.refGene	Chrom	Pos	Ref	Alts	Description
Estrogen and estrogen receptor	Nonsynonymous SNV	CYP1B1	chr2	38298404	C	T	The enzyme encoded by this gene localizes to the endoplasmic reticulum and metabolizes procarcinogens such as polycyclic aromatic hydrocarbons and 17 beta-estradiol.
	Frameshift insertion	CNOT2^	chr12	70747693	TAA	T, TA, TAAA	The protein encoded by this gene is involved in the negative regulation of intracellular estrogen receptor signaling pathway
Neuron development	Nonsynonymous SNV	AGRN	chr1	979016	G	A	The encoded protein contains several laminin G, Kazal-type serine protease inhibitor, and epidermal growth factor domains.
		CHRNA9	chr4	40351422	C	T	This protein is involved in cochlea hair cell development and is also expressed in the outer hair cells (OHCs) of the adult cochlea.
		OPRM1	chr6	154412447	C	G	This gene encodes one of at least three opioid receptors in humans: the mu opioid receptor (MOR). The MOR is the principal target of endogenous opioid peptides and opioid analgesic agents such as beta-endorphin and enkephalins.
		CTHRC1	chr8	104387987	G	A	This locus encodes a protein that may play a role in the cellular response to arterial injury through involvement in vascular remodeling.
		WASHC5	chr8	126093990	T	C	This gene encodes a 134 kDa protein named strumpellin that is predicted to have multiple transmembrane domains and a spectrin-repeat-containing domain. This ubiquitously expressed gene has its highest expression in skeletal muscle.
		NOTCH1	chr9	139412284	T	C	Notch signaling is an evolutionarily conserved intercellular signaling pathway that regulates interactions between physically adjacent cells through binding of Notch family receptors to their cognate ligands.
		PRKD1	chr14	30135343	A	G	The protein encoded by this gene is a serine/threonine protein kinase involved in many cellular processes, including cell migration and differentiation, MAPK8/JNK1 and Ras pathway signaling, MAPK1/3 (ERK1/2) pathway signaling, cell survival, and regulation of cell shape and adhesion.
		KCNJ14	chr19	48965151	G	A	This gene encodes a transmembrane protein that functions in copper transport across membranes. This protein is localized to the trans-Golgi network, where it is predicted to supply copper to copper-dependent enzymes in the secretory pathway.
		ATP7A&	chrX	77287057	G	A	This protein is localized to the trans-Golgi network, where it is predicted to supply copper to copper-dependent enzymes in the secretory pathway.
	Frameshift deletion	CCDC141	chr2	179733903	TTC	T	This gene is predicted to be involved in axon guidance and cell adhesion. It is also predicted to act upstream of or within centrosome localization and cerebral cortex radially oriented cell migration; to be located in the centrosome, cytoplasm, and plasma membrane; and to be active in neuron projections.
		IFT88	chr13	21179185	TACTA	T	This gene encodes a member of the tetratricopeptide repeat (TPR) family. The encoded protein is involved in cilium biogenesis.
	Frameshift insertion	FOXO6	chr1	41847881	CCA	CGGGACGCCCGCCTACA, C	It is predicted to enable DNA-binding transcription factor activity, RNA polymerase II-specific and RNA polymerase II cis-regulatory region sequence-specific DNA binding activity and predicted to be involved in positive regulation of dendritic spine development and regulation of transcription by RNA polymerase II.
		ANKS1A	chr6	35046418	A	AC	It is predicted to enable ephrin receptor binding activity and predicted to be involved in ephrin receptor signaling pathway, neuron remodeling, and substrate-dependent cell migration.
		ABR	chr17	934891	T	TG	The protein encoded by this gene contains a GTPase-activating protein domain, a domain found in members of the Rho family of GTP-binding proteins.
		FBXO7	chr22	32880039	C	CT	The F-box proteins constitute one of the four subunits of the ubiquitin protein ligase complex called SCFs (SKP1-cullin-F-box), which function in phosphorylation-dependent ubiquitination.
	Splicing	HNRNPR	chr1	23640196	C	CT	This gene encodes an RNA-binding protein that is a member of the spliceosome C complex, which functions in pre-mRNA processing and transport. The encoded protein also promotes transcription at the c-fos gene
		SHANK3	chr22	51135989	GTT	G	Shank proteins are multidomain scaffold proteins of the postsynaptic density that connect neurotransmitter receptors, ion channels, and other membrane proteins to the actin cytoskeleton and G-protein-coupled signaling pathways.

Discussion

Although susceptibility factors remain unclear, prior studies suggest a genetic component.¹⁵ Our study investigated potential genetic influences, focusing on ER-driven neurodevelopmental pathways in brain regions such as the ventromedial nucleus of the hypothalamus (VMN), medial preoptic area (mPOA), anteroventral periventricular nucleus (AVPV), and arcuate nucleus. These pathways are more active in males due to a pubertal testosterone surge, which is converted to estradiol and promotes male-typical behavior.^16,17 In contrast, limited ER activation in females leads to different developmental outcomes.¹⁶ In humans, gender-specific brain development is less understood, but parallels are seen in gender anxiety, which often worsens during adolescence. Prenatal androgen exposure, such as in females with congenital adrenal hyperplasia, is linked to more male-typical behavior and higher rates of gender anxiety.^18-21 Some males with gender anxiety are later diagnosed with congenital hypogonadotropic hypogonadism, characterized by low androgen levels.^17,22 Given the conservation of estrogen metabolism in mammals and the observed links between hormonal influence, neurodevelopment, and behavior, exploring related genetic variations may provide insights into the origins of gender anxiety and identity.

ER stimulation in VMN presynaptic neurons triggers glutamate release, activating postsynaptic N-methyl-D-aspartate (NMDA)receptors and downstream mitogen-activated protein (MAP) kinase pathways, leading to increased dendritic spine density—an effect linked to male-typical behavior in animal models.¹⁶ Four gene variants associated with this pathway were identified: ABR and ATP7A (involved in vesicle fusion and neuronal migration), Shank3 (regulates NMDA receptors),²³ and CHRNA9 (brain-expressed).²⁴ Given the differing susceptibilities between transgender males and females, it is essential to analyze them separately. ER pathway activation may have region- and sex-specific effects, suggesting genetic variants may influence susceptibility to assigned gender at birth. One notable variant is CD24, homozygous in four susceptible individuals; it is Y-linked and associated with testicular damage.^25,26

In addition to the previously mentioned candidate genes, WES identified several unique variants that do not fall within known sex-dimorphic neurodevelopmental pathways. Given that the genetic mechanisms underlying gender identity remain unclear, exploring these novel variants may uncover new neurodevelopmental pathways. These variants are not documented in the ExAC and dbSNP databases and were also absent in multiple nontransgender control samples, suggesting potential functional significance. Similarly, other studies using WES have identified rare variants not found in public databases, highlighting the effectiveness of WES in discovering novel genetic mutations.²⁷

Genetic research often focuses on identifying variants associated with disease, but it is important to recognize that not all traits are pathological. In 2019, the World Health Organization reclassified “gender identity disorder” as “gender incongruence,” moving it from the mental disorders chapter to the sexual health chapter in the ICD-10, acknowledging that transgender identity is not a mental illness.²²When investigating genetic factors related to gender identity, it’s crucial to use terminology that reflects this understanding. Terms like “pathogenic” or “mutation” may carry unintended negative connotations when applied to traits that are not disorders. Instead, researchers should focus on understanding the complex interplay of genetic, environmental, and social factors that contribute to gender identity, using terminology that respects the diversity and validity of transgender experiences. Gender identity is a multifaceted trait shaped by genetic, environmental, and social influences. The multigene threshold model posits that cumulative effects of multiple genetic variants contribute to the development of gender identity, allowing for a spectrum of phenotypes and challenging the notion that specific genetic variations can definitively predict an individual’s gender identity. Recent research utilizing WES has identified rare variants in genes associated with sexually dimorphic brain development, particularly within estrogen receptor–activated pathways.¹¹ These findings suggest a potential genetic contribution to gender incongruence and underscore the utility of WES in uncovering novel genetic variations.

Our study identified rare variants in 19 candidate genes potentially associated with gender development pathways in the brain. These variants were initially detected through WES and subsequently validated using Sanger sequencing. They were not present in the ExAC and dbSNP databases, or in multiple nontransgender controls, suggesting that they are not common polymorphisms. While these neurodevelopmental pathways have not yet been fully elucidated in humans as they have in animals, we propose that the genes involved in these pathways offer a viable avenue for investigating the genetic underpinnings of gender incongruence in humans.

Author contributions

Na Liu: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Jingyi Bai: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Nan Huang: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Yi Xu: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Xiangyun Long: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Xinyi Hu: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Jiaxin Wu: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Fei Liu: Data curation; Formal analysis; Investigation; Resources; Software; Visualization; Writing – original draft; Writing – review & editing. Zheng Lu: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Resources; Software; Visualization; Writing – original draft; Writing – review & editing

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of interest

The authors have no competing interests to declare that are relevant to the content of this article.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethical approval

All the procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study was approved by the Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine (no. 2021-76).

Consent to participate

Informed consent was obtained from all the individual participants included in the study.