A genome‐wide association study of post‐traumatic stress disorder in traumatised Chinese populations

To the editor, Genome-wide association studies (GWASs) of post-traumatic stress disorder (PTSD) have been extensively performed in populations of European and African ancestry1 but remain largely unexplored for Chinese populations. We report the first PTSD GWAS in two independent Chinese cohorts comprising 1223 trauma-exposed individuals (288 cases and 935 controls). Gene-set analysis was performed to explore the biological function and the mechanism of PTSD. Genome-wide association analysis identified a locus at 15q26. 2 (LINC00923, rs11854251) approaching genome-wide significance (p = 3. 88 × 10−8). This association was not observed in prior PTSD GWASs of European and other populations. 1 Pathway enrichment analysis indicates that betaine metabolism and carnitine metabolism are associated with PTSD. This study provides new insights into the genetic architecture and pathophysiological mechanisms underlying PTSD susceptibility in Chinese populations. PTSD is a mental disorder that occurs after exposure to traumatic events; its symptoms include intrusive re-experiencing, avoidance and hyperarousal. 2 PTSD exhibits a complex aetiology, influenced by both genetic and environmental factors. The pathogenesis of PTSD is still not well understood, and evidence indicates that PTSD is typically polygenic, arising from thousands of genetic variations distributed widely across the genome. 3 GWAS has provided important insights into the genetic risk factors and potential mechanisms of PTSD. A large-scale meta-analysis study reported 95 genome-wide significant loci. 1 However, thus far, GWASs of PTSD have almost entirely been carried out in populations of European and African ancestry, 1 with no study to date carried out in a Chinese population. China, situated in the Asia-Pacific region, has a huge population of about 1. 4 billion. Affected by geological activity and, more recently, by global climate change, China has experienced a high frequency of natural disasters throughout its history. A survey of natural disasters in China by the Office of the National Disaster Prevention reported that in 2024 alone, natural disasters (e. g. , earthquakes, floods, typhoons and wildfires) affected approximately 6. 7% (94 million) of the population. Furthermore, the National Report on Injury Prevention in China (2007) reported an alarming frequency of unintentional injuries (e. g. , traffic accidents, falls, poisoning and drowning). These widespread traumatic exposures represent a significant risk factor for the development of PTSD. Genetic studies of PTSD are highly heterogeneous due to differences in environment, ethnicity, trauma type, and so on. A PTSD GWAS carried out in a Chinese population would be highly likely to discover novel susceptibility genes. Thus, such a study is essential if PTSD's genetic landscape is to be comprehensively mapped. The two cohorts examined in this study are the Deyang Hanwang and the Deyang Hospital cohorts. The Deyang Hanwang cohort comprised 1035 subjects who had personally experienced the 2008 Wenchuan earthquake. Exposure to trauma was quantified using a 10-item questionnaire. 4 The Deyang Hospital cohort comprised 481 treatment-naive patients recruited from the Psychosomatic Medicine Department of Deyang People's Hospital. The psychometric properties of the Posttraumatic Diagnostic Scale for Diagnostic and Statistical Manual of Mental Disorders (DSM-5) 5 were applied to screen for potentially traumatic events and identify index traumatic events for measuring PTSD symptoms. An overview of the demographic characteristics and traumatic event situations of the two GWAS cohorts is shown in table S1. For both cohorts, all control participants were exposed to trauma. PTSD was diagnosed according to the criteria in the fifth edition of the DSM-5. 2 PTSD symptoms were measured using the Posttraumatic Stress Disorder Checklist for DSM-5. 6 Peripheral blood was collected, and genomic DNA was extracted from all participants. Genotyping was performed using an Illumina ASAMD-24v1-0₂0025976A1 single nucleotide polymorphism (SNP) array (743. 7k SNPs). Following quality control, approximately 495. 8k SNPs remained, representing 1223 individuals, of whom 840 were Deyang Hanwang subjects (130 cases and 710 controls) and 383 were Deyang Hospital cohort subjects (158 cases and 225 controls). Approximately 4087. 2k genotyped and imputed SNPs were analysed. An association test was performed using an additive model in a logistic regression model. The covariates were the first principal component of population stratification (PC1), sex, age and trauma exposure for the Deyang Hanwang cohort and PC1, sex and age for the Deyang Hospital cohort. A meta-analysis using the fixed-effect inverse variance model was used to combine the results from the two cohorts. The I2 index and Cochran's Q test were used to test heterogeneity. The criterion p 18 years old, i. e. , child vs adult) in the Deyang Hospital cohort, and the Deyang Hanwang cohort exhibits disproportionate female representation (cases: 77. 7% and controls: 63. 9%) (see table S1). Thus, the developmental stage might be a confounding factor, and sex may play a crucial role in PTSD development. 17 To explore the potential confounding effects of age and sex, additional stratified, sensitive, and interaction analyses were performed for the genome-wide significance of SNP rs11854251. For the Deyang Hospital cohort, samples were stratified by age (≤ 18 and > 18 years) for separate analysis, and the results were combined in a meta-analysis. For the Deyang Hanwang cohort, we performed a sensitivity analysis (comparing deviance and coefficients from the regression model with and without the covariate sex), sex–SNP interaction tests using logistic regression, and sex-stratified analyses to assess potential gender-specific effects. The results show that age and sex did not affect our results (see tables S8 and S9). For the Deyang Hospital cohort, there was no heterogeneity between the two age groups (Cochran's Q test, p > 0. 1). For the Deyang Hanwang cohort, the absence of covariate sex led to only a 2% increase in regression model deviance and only a 2% decrease in coefficients; there was no interaction effect between sex and gene (pinteraction = 0. 832), and the two sex groups exhibited no heterogeneity (Cochran's Q test p > 0. 1). Regarding age discrepancy, the mean age differed between cases and controls in both cohorts (Deyang Hanwang cohort: 54. 70 vs 50. 72 and Deyang Hospital cohort: 25. 20 vs 34. 56). Age was thus added as a covariate to the association test to overcome this limitation. The sample size in our study was small and only included a single population. The association analysis of our GWAS has low statistical power due to the small sample size. This potentially may lead to a limited statistical significance for the top SNP. In fact, the p value of the top SNP is just slightly below the genome-wide significance level. In competitive gene-set analysis, statistical power depends strongly on phenotype heritability, and increasing sample size provides only limited gains due to the approach's structure. 18 Thus, the statistical power of the pathway analysis could not be very low and might not have much effect on the statistical significance of the discovered metabolic pathways. In particular, the p value (3. 88 × 10−8) for rs11854251 is extremely close to the genome-wide significance threshold (p = 5 × 10−8). So, validation in an independent cohort is required in the future. In addition, future research should use a larger sample size and explore a broader spectrum of genetic variation, including other variant types, epigenetic factors and cross-ethnic validation. Diverse functional genomics approaches should be employed to achieve a comprehensive, integrative understanding of PTSD. In conclusion, this paper reports the first PTSD GWAS in a Chinese population. It identifies a novel, genome-wide significant locus, LINC00923 at 15q26. 2, with population-specific associations in Chinese individuals, thereby advancing both ethnic-specific risk prediction and therapeutic target discovery. Kunlin Zhang, Mingyue Gao and Haibo Yang analysed data. Chengqi Cao, Ping Liu and Yajie Bi collected samples. Yuwei Han and Ling Xu performed the experiment. Kunlin Zhang and Mingyue Gao wrote the paper. Li Wang and Kunlin Zhang obtained fundings. Li Wang and the This work was by Project of the National of China and the National of China and The no of This study has been by the of the of Chinese of and was in with and the of participants or their Kunlin Zhang, is an of the of and Mental of Chinese of in obtained a and a both from China, in and did a at the and the in from and work at in is a of the Mental of the China Association for Mental on data on mental research include and biological mechanisms of mental disorders by statistical and The is not for the or of information by the than should be to the corresponding for the