Transcriptomic Imputation Research Articles

Transcriptomic imputation identifies tissue-specific genes associated with cervical myelopathy

Degenerative cervical myelopathy (DCM) is a progressive spinal condition that can lead to severe neurological dysfunction. Despite its degenerative pathophysiology, family history has shown to be a largely important factor in incidence and progression, suggesting that inherent genetic predisposition may play a role in pathophysiology. To determine the tissue-specific, functional genetic basis of hereditary predisposition to cervical myelopathy. Retrospective case-control study using patient genetics and matched EHR from the Mount Sinai BioMe Biobank. In a large, diverse, urban biobank of 32,031 individuals, with 558 individuals with cervical myopathy, we applied transcriptomic imputation to identify genetically regulated gene expression signatures associated with DCM. We performed drug-repurposing analysis using the CMAP database to identify candidate therapeutic interventions to reverse the cervical myelopathy-associated gene signature. We identified 16 genes significantly associated with DCM across five different tissues, suggesting tissue-specific manifestations of inherited genetic risk (upregulated: HES6, PI16, TMEM183A, BDH2, LINC00937, CLEC4D, USP43, SPATA1; downregulated: TTC12, CDK5, PAFAH1B2, RCSD1, KLHL29, PTPRG, RP11-620J15.3, C1RL). Drug repurposing identified 22 compounds with the potential to reverse the DCM-associated signature, suggesting points of therapeutic intervention. The inherited genetic risk for cervical myelopathy is functionally associated with genes involved in tissue-specific nociceptive and proliferative processes. These signatures may be reversed by candidate therapeutics with nociceptive, calcium channel modulating, and antiproliferative effects. Understanding the genetic basis of DCM provides critical insights into the hereditary factors contributing to the disease, allowing for more personalized and targeted therapeutic approaches. The identification of candidate drugs through transcriptomic imputation and drug repurposing analysis offers potential new treatments that could significantly improve patient outcomes and quality of life by addressing the underlying genetic mechanisms of DCM.

A genetic study to identify pathogenic mechanisms and drug targets for benign prostatic hyperplasia: a multi-omics Mendelian randomization study

Benign prostatic hyperplasia (BPH) as a common geriatric disease in urology, the incidence and prevalence are rapidly increasing with the aging society, prompting an urgent need for effective prevention and treatment of BPH. However, limited therapeutic efficacy and higher risk of complications result in the treatment of BPH remaining challenging. The unclear pathogenic mechanism also hampers further exploration of therapeutic approaches for BPH. In this study, we used multi-omics methods to integrate genomics, transcriptomics, immunomics, and metabolomics data and identify biomolecules associated with BPH. We performed transcriptomic imputation, summary data-based Mendelian randomization (SMR), joint/conditional analysis, colocalization analysis, and FOCUS to explore high-confidence genes associated with BPH in blood and prostate tissue. Subsequently, three-step SMR was used to identify the DNA methylation sites regulating high-confidence genes to improve the pathogenic pathways of BPH. We also used cis-instruments of druggable genes to conduct SMR analysis to find potential drug targets for BPH. Finally, we used MR analysis to explore the immune pathways and metabolomics related to BPH. Multiple analytical methods identified BTN3A2 (Blood: TWAS Z score = 5.02912, TWAS P = 4.93 × 10–7; Prostate: TWAS Z score = 4.89, TWAS P = 1.01 × 10–6) and C4A (Blood: TWAS Z score = 4.90754, TWAS P = 9.22 × 10–7; Prostate: TWAS Z score = 5.084, TWAS P = 3.70 × 10–7) as high-confidence genes for BPH and identified the cg14345882-BTN3A2-BPH pathogenic pathway. We also used druggable gene data to identify 30 promising therapeutic target genes, including BTN3A2 and C4A. For MR analysis of immune pathways, we identified immune cell surface molecules as well as the inflammatory factor IL-17 (OR = 1.25, 95% CI = 1.09–1.43, PFDR = 0.12, Maximum likelihood) as risk factors for BPH. In addition, we found that disulfide levels of cysteinylglycine (OR = 1.11, 95% CI = 1.05–1.18, P = 5.18 × 10–4, Weighted median), oxidation levels of cysteinylglycine (OR = 1.09, 95% CI = 1.04–1.14, P = 3.87 × 10–4, Weighted median), and sebacate levels (OR = 1.05, 95% CI = 1.02–1.08, P = 3.0 × 10–4, Maximum likelihood) increase the risk of BPH. This multi-omics study explored biomolecules associated with BPH, improved the pathogenic pathways of BPH, and identified promising therapeutic targets. Our results provide evidence for future studies aimed at developing appropriate therapeutic interventions.

Major Psychiatric Disorders, Substance Use Behaviors, and Longevity

Observational studies suggest that major psychiatric disorders and substance use behaviors reduce longevity, making it difficult to disentangle their relationships with aging-related outcomes. To evaluate the associations between the genetic liabilities for major psychiatric disorders, substance use behaviors (smoking and alcohol consumption), and longevity. This 2-sample mendelian randomization (MR) study assessed associations between psychiatric disorders, substance use behaviors, and longevity using single-variable and multivariable models. Multiomics analyses were performed elucidating transcriptomic underpinnings of the MR associations and identifying potential proteomic therapeutic targets. This study sourced summary-level genome-wide association study (GWAS) data, gene expression, and proteomic data from cohorts of European ancestry. Analyses were performed from May 2022 to November 2023. Genetic susceptibility for major depression (n = 500 199), bipolar disorder (n = 413 466), schizophrenia (n = 127 906), problematic alcohol use (n = 435 563), weekly alcohol consumption (n = 666 978), and lifetime smoking index (n = 462 690). The main outcome encompassed aspects of health span, lifespan, and exceptional longevity. Additional outcomes were epigenetic age acceleration (EAA) clocks. Findings from multivariable MR models simultaneously assessing psychiatric disorders and substance use behaviorsm suggest a negative association between smoking and longevity in cohorts of European ancestry (n = 709 709; 431 503 [60.8%] female; β, -0.33; 95% CI, -0.38 to -0.28; P = 4.59 × 10-34) and with increased EAA (n = 34 449; 18 017 [52.3%] female; eg, PhenoAge: β, 1.76; 95% CI, 0.72 to 2.79; P = 8.83 × 10-4). Transcriptomic imputation and colocalization identified 249 genes associated with smoking, including 36 novel genes not captured by the original smoking GWAS. Enriched pathways included chromatin remodeling and telomere assembly and maintenance. The transcriptome-wide signature of smoking was inversely associated with longevity, and estimates of individual smoking-associated genes, eg, XRCC3 and PRMT6, aligned with the smoking-longevity MR analyses, suggesting underlying transcriptomic mediators. Cis-instrument MR prioritized brain proteins associated with smoking behavior, including LY6H (β, 0.02; 95% CI, 0.01 to 0.03; P = 2.37 × 10-6) and RIT2 (β, 0.02; 95% CI, 0.01 to 0.03; P = 1.05 × 10-5), which had favorable adverse-effect profiles across 367 traits evaluated in phenome-wide MR. The findings suggest that the genetic liability of smoking, but not of psychiatric disorders, is associated with longevity. Transcriptomic associations offer insights into smoking-related pathways, and identified proteomic targets may inform therapeutic development for smoking cessation strategies.

Transcriptomic imputation of genetic risk variants uncovers novel whole-blood biomarkers of Parkinson’s disease

Blood-based gene expression signatures could potentially be used as biomarkers for PD. However, it is unclear whether genetically-regulated transcriptomic signatures can provide novel gene candidates for use as PD biomarkers. We leveraged on the Genotype-Tissue Expression (GTEx) database to impute whole-blood transcriptomic expression using summary statistics of three large-scale PD GWAS. A random forest classifier was used with the consensus whole-blood imputed gene signature (IGS) to discriminate between cases and controls. Outcome measures included Area under the Curve (AUC) of Receiver Operating Characteristic (ROC) Curve. We demonstrated that the IGS (n = 37 genes) is conserved across PD GWAS studies and brain tissues. IGS discriminated between cases and controls in an independent whole-blood RNA-sequencing study (1176 PD, 254 prodromal, and 860 healthy controls) with mean AUC and accuracy of 64.8% and 69.4% for PD cohort, and 78.8% and 74% for prodromal cohort. PATL2 was the top-performing imputed gene in both PD and prodromal PD cohorts, whose classifier performance varied with biological sex (higher performance for males and females in the PD and prodromal PD, respectively). Single-cell RNA-sequencing studies (scRNA-seq) of healthy humans and PD patients found PATL2 to be enriched in terminal effector CD8+ and cytotoxic CD4+ cells, whose proportions are both increased in PD patients. We demonstrated the utility of GWAS transcriptomic imputation in identifying novel whole-blood transcriptomic signatures which could be leveraged upon for PD biomarker derivation. We identified PATL2 as a potential biomarker in both clinical and prodromic PD.

Genetically Regulated Gene Expression in the Brain Associated With Chronic Pain: Relationships With Clinical Traits and Potential for Drug Repurposing

BackgroundChronic pain is a common, poorly understood condition. Genetic studies including genome-wide association studies have identified many relevant variants, which have yet to be translated into full understanding of chronic pain. Transcriptome-wide association studies using transcriptomic imputation methods such as S-PrediXcan can help bridge this genotype-phenotype gap. MethodsWe carried out transcriptomic imputation using S-PrediXcan to identify genetically regulated gene expression associated with multisite chronic pain in 13 brain tissues and whole blood. Then, we imputed genetically regulated gene expression for over 31,000 Mount Sinai BioMe participants and performed a phenome-wide association study to investigate clinical relationships in chronic pain–associated gene expression changes. ResultsWe identified 95 experiment-wide significant gene-tissue associations (p < 7.97 × 10−7), including 36 unique genes and an additional 134 gene-tissue associations reaching within-tissue significance, including 53 additional unique genes. Of the 89 unique genes in total, 59 were novel for multisite chronic pain and 18 are established drug targets. Chronic pain genetically regulated gene expression for 10 unique genes was significantly associated with cardiac dysrhythmia, metabolic syndrome, disc disorders/dorsopathies, joint/ligament sprain, anemias, and neurologic disorder phecodes. Phenome-wide association study analyses adjusting for mean pain score showed that associations were not driven by mean pain score. ConclusionsWe carried out the largest transcriptomic imputation study of any chronic pain trait to date. Results highlight potential causal genes in chronic pain development and tissue and direction of effect. Several gene results were also drug targets. Phenome-wide association study results showed significant associations for phecodes including cardiac dysrhythmia and metabolic syndrome, thereby indicating potential shared mechanisms.

Multivariate genome-wide analysis of aging-related traits identifies novel loci and new drug targets for healthy aging.

The concept of aging is complex, including many related phenotypes such as healthspan, lifespan, extreme longevity, frailty and epigenetic aging, suggesting shared biological underpinnings; however, aging-related endpoints have been primarily assessed individually. Using data from these traits and multivariate genome-wide association study methods, we modeled their underlying genetic factor ('mvAge'). mvAge (effective n = ~1.9 million participants of European ancestry) identified 52 independent variants in 38 genomic loci. Twenty variants were novel (not reported in input genome-wide association studies). Transcriptomic imputation identified age-relevant genes, including VEGFA and PHB1. Drug-target Mendelian randomization with metformin target genes showed a beneficial impact on mvAge (P value = 8.41 × 10-5). Similarly, genetically proxied thiazolidinediones (P value = 3.50 × 10-10), proprotein convertase subtilisin/kexin 9 inhibition (P value = 1.62 × 10-6), angiopoietin-like protein 4, beta blockers and calcium channel blockers also had beneficial Mendelian randomization estimates. Extending the drug-target Mendelian randomization framework to 3,947 protein-coding genes prioritized 122 targets. Together, these findings will inform future studies aimed at improving healthy aging.

Multi-omic underpinnings of epigenetic aging and human longevity

Biological aging is accompanied by increasing morbidity, mortality, and healthcare costs; however, its molecular mechanisms are poorly understood. Here, we use multi-omic methods to integrate genomic, transcriptomic, and metabolomic data and identify biological associations with four measures of epigenetic age acceleration and a human longevity phenotype comprising healthspan, lifespan, and exceptional longevity (multivariate longevity). Using transcriptomic imputation, fine-mapping, and conditional analysis, we identify 22 high confidence associations with epigenetic age acceleration and seven with multivariate longevity. FLOT1, KPNA4, and TMX2 are novel, high confidence genes associated with epigenetic age acceleration. In parallel, cis-instrument Mendelian randomization of the druggable genome associates TPMT and NHLRC1 with epigenetic aging, supporting transcriptomic imputation findings. Metabolomics Mendelian randomization identifies a negative effect of non-high-density lipoprotein cholesterol and associated lipoproteins on multivariate longevity, but not epigenetic age acceleration. Finally, cell-type enrichment analysis implicates immune cells and precursors in epigenetic age acceleration and, more modestly, multivariate longevity. Follow-up Mendelian randomization of immune cell traits suggests lymphocyte subpopulations and lymphocytic surface molecules affect multivariate longevity and epigenetic age acceleration. Our results highlight druggable targets and biological pathways involved in aging and facilitate multi-omic comparisons of epigenetic clocks and human longevity.

Mapping anorexia nervosa genes to clinical phenotypes.

Anorexia nervosa (AN) is a psychiatric disorder with complex etiology, with a significant portion of disease risk imparted by genetics. Traditional genome-wide association studies (GWAS) produce principal evidence for the association of genetic variants with disease. Transcriptomic imputation (TI) allows for the translation of those variants into regulatory mechanisms, which can then be used to assess the functional outcome of genetically regulated gene expression (GReX) in a broader setting through the use of phenome-wide association studies (pheWASs) in large and diverse clinical biobank populations with electronic health record phenotypes. Here, we applied TI using S-PrediXcan to translate the most recent PGC-ED AN GWAS findings into AN-GReX. For significant genes, we imputed AN-GReX in the Mount Sinai BioMe™ Biobank and performed pheWASs on over 2000 outcomes to test the clinical consequences of aberrant expression of these genes. We performed a secondary analysis to assess the impact of body mass index (BMI) and sex on AN-GReX clinical associations. Our S-PrediXcan analysis identified 53 genes associated with AN, including what is, to our knowledge, the first-genetic association of AN with the major histocompatibility complex. AN-GReX was associated with autoimmune, metabolic, and gastrointestinal diagnoses in our biobank cohort, as well as measures of cholesterol, medications, substance use, and pain. Additionally, our analyses showed moderation of AN-GReX associations with measures of cholesterol and substance use by BMI, and moderation of AN-GReX associations with celiac disease by sex. Our BMI-stratified results provide potential avenues of functional mechanism for AN-genes to investigate further.

Integration of genetic, transcriptomic, and clinical data provides insight into 16p11.2 and 22q11.2 CNV genes

BackgroundDeletions and duplications of the multigenic 16p11.2 and 22q11.2 copy number variant (CNV) regions are associated with brain-related disorders including schizophrenia, intellectual disability, obesity, bipolar disorder, and autism spectrum disorder (ASD). The contribution of individual CNV genes to each of these identified phenotypes is unknown, as well as the contribution of these CNV genes to other potentially subtler health implications for carriers. Hypothesizing that DNA copy number exerts most effects via impacts on RNA expression, we attempted a novel in silico fine-mapping approach in non-CNV carriers using both GWAS and biobank data.MethodsWe first asked whether gene expression level in any individual gene in the CNV region alters risk for a known CNV-associated behavioral phenotype(s). Using transcriptomic imputation, we performed association testing for CNV genes within large genotyped cohorts for schizophrenia, IQ, BMI, bipolar disorder, and ASD. Second, we used a biobank containing electronic health data to compare the medical phenome of CNV carriers to controls within 700,000 individuals in order to investigate the full spectrum of health effects of the CNVs. Third, we used genotypes for over 48,000 individuals within the biobank to perform phenome-wide association studies between imputed expressions of individual 16p11.2 and 22q11.2 genes and over 1500 health traits.ResultsUsing large genotyped cohorts, we found individual genes within 16p11.2 associated with schizophrenia (TMEM219, INO80E, YPEL3), BMI (TMEM219, SPN, TAOK2, INO80E), and IQ (SPN), using conditional analysis to identify upregulation of INO80E as the driver of schizophrenia, and downregulation of SPN and INO80E as increasing BMI. We identified both novel and previously observed over-represented traits within the electronic health records of 16p11.2 and 22q11.2 CNV carriers. In the phenome-wide association study, we found seventeen significant gene-trait pairs, including psychosis (NPIPB11, SLX1B) and mood disorders (SCARF2), and overall enrichment of mental traits.ConclusionsOur results demonstrate how integration of genetic and clinical data aids in understanding CNV gene function and implicates pleiotropy and multigenicity in CNV biology.

Testing Glucocorticoid Mediated Differential Gene Expression in IPSC Derived Neural Cultures as a Model for PTSD

Posttraumatic Stress Disorder (PTSD) is a common complex psychiatric disorder affecting 6.1% of Americans, and no efficacious cellular models exists for studying the disease. Utilizing a transcriptomic imputation approach with the latest GWAS and TWAS analyses of PTSD, our group has recently predicted DLPFC gene expression, and its constituent cell types, in the disease's pathology.

Fine-mapping and QTL tissue-sharing information improves the reliability of causal gene identification.

The integration of transcriptomic studies and genome‐wide association studies (GWAS) via imputed expression has seen extensive application in recent years, enabling the functional characterization and causal gene prioritization of GWAS loci. However, the techniques for imputing transcriptomic traits from DNA variation remain underdeveloped. Furthermore, associations found when linking eQTL studies to complex traits through methods like PrediXcan can lead to false positives due to linkage disequilibrium between distinct causal variants. Therefore, the best prediction performance models may not necessarily lead to more reliable causal gene discovery. With the goal of improving discoveries without increasing false positives, we develop and compare multiple transcriptomic imputation approaches using the most recent GTEx release of expression and splicing data on 17,382 RNA‐sequencing samples from 948 post‐mortem donors in 54 tissues. We find that informing prediction models with posterior causal probability from fine‐mapping (dap‐g) and borrowing information across tissues (mashr) can lead to better performance in terms of number and proportion of significant associations that are colocalized and the proportion of silver standard genes identified as indicated by precision‐recall and receiver operating characteristic curves. All prediction models are made publicly available at predictdb.org.

Analysis of Genetically Regulated Gene Expression Identifies a Prefrontal PTSD Gene, SNRNP35, Specific to Military Cohorts

To reveal post-traumatic stress disorder (PTSD) genetic risk influences on tissue-specific gene expression, we use brain and non-brain transcriptomic imputation. We impute genetically regulated gene expression (GReX) in 29,539 PTSD cases and 166,145 controls from 70 ancestry-specific cohorts and identify 18 significant GReX-PTSD associations corresponding to specific tissue-gene pairs. The results suggest substantial genetic heterogeneity based on ancestry, cohort type (military versus civilian), and sex. Two study-wide significant PTSD associations are identified in European and military European cohorts; ZNF140 is predicted to be upregulated in whole blood, and SNRNP35 is predicted to be downregulated in dorsolateral prefrontal cortex, respectively. In peripheral leukocytes from 175 marines, the observed PTSD differential gene expression correlates with the predicted differences for these individuals, and deployment stress produces glucocorticoid-regulated expression changes that include downregulation of both ZNF140 and SNRNP35. SNRNP35 knockdown in cells validates its functional role in U12-intron splicing. Finally, exogenous glucocorticoids in mice downregulate prefrontal Snrnp35 expression.

Pairwise common variant meta-analyses of schizophrenia with other psychiatric disorders reveals shared and distinct gene and gene-set associations

The complex aetiology of schizophrenia is postulated to share components with other psychiatric disorders. We investigated pleiotropy amongst the common variant genomics of schizophrenia and seven other psychiatric disorders using a multimarker association test. Transcriptomic imputation was then leveraged to investigate the functional significance of variation mapped to these genes, prioritising several interesting functional candidates. Gene-based analysis of common variation revealed 67 schizophrenia-associated genes shared with other psychiatric phenotypes, including bipolar disorder, major depressive disorder, ADHD and autism-spectrum disorder. In addition, we uncovered 78 genes significantly enriched with common variant associations for schizophrenia that were not linked to any of these seven disorders (P > 0.05). Multivariable gene-set association suggested that common variation enrichment within biologically constrained genes observed for schizophrenia also occurs across several psychiatric phenotypes. Pairwise meta-analysis of schizophrenia and each psychiatric phenotype was implemented and identified 330 significantly associated genes (PMeta < 2.7 × 10−6) that were only nominally associated with each disorder individually (P < 0.05). These analyses consolidate the overlap between the genomic architecture of schizophrenia and other psychiatric disorders, uncovering several candidate pleiotropic genes which warrant further investigation.

TIGAR: An Improved Bayesian Tool for Transcriptomic Data Imputation Enhances Gene Mapping of Complex Traits

The transcriptome-wide association studies (TWASs) that test for association between the study trait and the imputed gene expression levels from cis-acting expression quantitative trait loci (cis-eQTL) genotypes have successfully enhanced the discovery of genetic risk loci for complex traits. By using the gene expression imputation models fitted from reference datasets that have both genetic and transcriptomic data, TWASs facilitate gene-based tests with GWAS data while accounting for the reference transcriptomic data. The existing TWAS tools like PrediXcan and FUSION use parametric imputation models that have limitations for modeling the complex genetic architecture of transcriptomic data. Therefore, to improve on this, we employ a nonparametric Bayesian method that was originally proposed for genetic prediction of complex traits, which assumes a data-driven nonparametric prior for cis-eQTL effect sizes. The nonparametric Bayesian method is flexible and general because it includes both of the parametric imputation models used by PrediXcan and FUSION as special cases. Our simulation studies showed that the nonparametric Bayesian model improved both imputation R2 for transcriptomic data and the TWAS power over PrediXcan when ≥1% cis-SNPs co-regulate gene expression and gene expression heritability ≤0.2. In real applications, the nonparametric Bayesian method fitted transcriptomic imputation models for 57.8% more genes over PrediXcan, thus improving the power of follow-up TWASs. We implement both parametric PrediXcan and nonparametric Bayesian methods in a convenient software tool "TIGAR" (Transcriptome-Integrated Genetic Association Resource), which imputes transcriptomic data and performs subsequent TWASs using individual-level or summary-level GWAS data.

200. Dissecting the Transcriptomic and Phenotypic Complexity of PTSD With Transcriptomic Imputation and Bayesian Machine Learning

PTSD occurs in some trauma-exposed individuals, but its molecular basis remains uncharacterized. Previously, we have associated genetically regulated gene expression across tissues with PTSD (Huckins et al. 2018). Here, we study the relationship between genetically-regulated transcriptional variation, and phenotypic variation in PTSD, in a cohort of 9400 inner-city, low-socioeconomic-status, primarily-African-American patients of the Grady Memorial Hospital.

Gene expression imputation across multiple brain regions provides insights into schizophrenia risk.

Transcriptomic imputation approaches combine eQTL reference panels with large-scale genotype data in order to test associations between disease and gene expression. These genic associations could elucidate signals in complex genome-wide association study (GWAS) loci and may disentangle the role of different tissues in disease development. We used the largest eQTL reference panel for the dorso-lateral prefrontal cortex (DLPFC) to create a set of gene expression predictors and demonstrate their utility. We applied DLPFC and 12 GTEx-brain predictors to 40,299 schizophrenia cases and 65,264 matched controls for a large transcriptomic imputation study of schizophrenia. We identified 413 genic associations across 13 brain regions. Stepwise conditioning identified 67 non-MHC genes, of which 14 did not fall within previous GWAS loci. We identified 36 significantly enriched pathways, including hexosaminidase-A deficiency, and multiple porphyric disorder pathways. We investigated developmental expression patterns among the 67 non-MHC genes and identified specific groups of pre- and postnatal expression.

Analysis of Genetically Regulated Gene Expression Identifies a Trauma Type Specific PTSD Gene, SNRNP35

PTSD has significant genetic heritability; however, it is unclear how genetic risk influences tissue-specific gene expression. We used brain and non-brain transcriptomic imputation models to impute genetically regulated gene expression (GReX) in 9,087 PTSD-cases and 23,811 controls and identified thirteen significant GReX-PTSD associations. The results suggest substantial genetic heterogeneity between civilian and military PTSD cohorts. The top study-wide significant PTSD-association was with predicted downregulation of the Small Nuclear Ribonucleoprotein U11/U12 Subunit 35 (SNRNP35) in the BA9 region of the prefrontal cortex (PFC) in military cohorts. In peripheral leukocytes from 175 U.S. Marines, the observed PTSD differential gene expression correlated with the predicted blood GReX differences for these individuals, and deployment stress downregulated SNRNP35 expression, primarily in Marines with post-deployment PTSD. SNRNP35 is a subunit of the minor spliceosome complex and SNRNP35 knockdown in cells validated its functional importance in U12-intron splicing. Finally, mimicking acute activation of the endogenous stress axis in mice downregulated PFC Snrnp35 expression.

63 - TRANSCRIPTOMIC IMPUTATION ANALYSIS IN ANOREXIA NERVOSA IDENTIFIES BOTH METABOLIC AND PSYCHIATRIC AETIOLOGIES

Background Anorexia Nervosa (AN) is a complex and serious eating disorder, occurring in ~1% of individuals. Despite having the highest mortality rate of any psychiatric disorder, little is known about the aetiology of the AN, and few effective treatments exist. Global efforts to collect large samples from individuals with AN have been highly successful, and a recent study consequently identified the first genome-wide significant locus involved in AN, with significant genetic correlations with multiple metabolic indices. This result, coupled with other recent studies and epidemiological evidence, implies that previous characterizations of AN as a purely psychiatric disorder should be revisited. Rather, it seems that both psychiatric and metabolic systems may be involved. Methods In order to elucidate more of the system-specific aetiology of AN, we applied transcriptomic imputation methods to 3,495 cases and 10,982 controls, collected by the PGC-ED working group. Transcriptomic Imputation (TI) methods approaches use machine-learning methods to impute tissue-specific gene expression from large genotype data using curated eQTL reference panels. These offer an exciting opportunity to compare gene associations across neurological and metabolic tissues. Here, we applied CommonMind Consortium (CMC) and GTEx-derived gene expression prediction models for 12 brain tissues and 12 tissues with potential metabolic involvement (Adipose, Adrenal Gland, 2 Colon, 3 Esophagus, Liver, Pancreas, Small Intestine, Spleen, Stomach). In total, these analyses include 234,968 gene-tissue association tests. Results We identified 35 significant gene-tissue associations within the large chromosome 12 region described in the recent PGC-ED GWAS. We applied a forward-stepwise conditional analysis analogous to GCTA-CoJo in order to identify independent associations within this region and across tissues, and identified two independently associated genes, both in metabolic tissues; RPS26, in the Spleen (p=2.59e-08), and C12orf49, in the Thyroid (p=4.01e-06). We also identified two further regions with genome-wide significant gene-tissue associations, both in brain tissues; REEP5, in the DLPFC (p=8.52e-07), and CUL3, in the Caudate Basal Ganglia (p=1.8e-06). Discussion Our results support a model of AN risk influenced by both metabolic and psychiatric factors. We plan to follow these analyses with investigations of the overlap between AN and other psychiatric and metabolic phenotypes, as well as analysis of potentially implicated molecular pathways.

Analysis of Genetically Regulated Gene Expression Identifies a Prefrontal PTSD Gene, &lt;i&gt;SNRNP35&lt;/i&gt;, Specific to Military Cohorts

PTSD has significant genetic heritability; however, it is unclear how genetic risk influences tissue-specific gene expression. We used brain and non-brain transcriptomic imputation models to impute genetically regulated gene expression (GReX) in 29,539 PTSD-cases and 166,145 controls from 70 unique PTSD cohorts and identified eighteen significant GReX-PTSD associations. The results suggest substantial genetic heterogeneity based on ancestry, cohort-type (military vs. civilian) and sex. The two study-wide significant PTSD-associations were identified in European (23195 cases/174642 controls) and military European cohorts (6004 cases/27538 controls); ZNF140 was predicted to be upregulated in whole blood and SNRNP35 was predicted to be downregulated in the BA9 region of the prefrontal cortex (PFC), respectively. Both associations were stronger in men compared to women. In peripheral leukocytes from 175 U.S. Marines, the observed PTSD differential gene expression correlated with the predicted blood GReX differences for these individuals, and deployment stress produced glucocorticoid- regulated expression changes that included downregulation of both ZNF140 and SNRNP35 expression. SNRNP35 is a subunit of the minor spliceosome complex and SNRNP35 knockdown in cells validated its functional importance in U12-intron splicing. Finally, mimicking acute activation of the endogenous stress axis with exogenous glucocorticoids in mice downregulated PFC Snrnp35 expression.

SU11 - GENETICALLY PREDICTED GENE EXPRESSION IN THE BRAIN AND PERIPHERAL TISSUES ASSOCIATES WITH PTSD

Background Post-Traumatic Stress Disorder (PTSD) is a debilitating psychiatric disorder occurring in individuals exposed to trauma. To date, little is known about the genetic aetiology of the disorder, although the latest GWAS (carried out by the PGC-PTSD working group) demonstrates that genetic heritability is in line with other psychiatric disorders. PTSD development involves multi-systemic dysregulation in multiple brain regions and diverse peripheral tissues. Some peripheral systems are particularly interesting since epidemiologic evidence suggests that PTSD patients commonly have cardiovascular, metabolic and immune dysregulation. Methods Transcriptomic Imputation approaches use machine-learning methods to impute gene expression from large genotype data using curated eQTL reference panels. These offer an exciting opportunity to compare gene associations across neurological and peripheral tissues. Here, we apply CommonMind Consortium (CMC) and Genotype-Tissue Expression (GTEx) derived gene expression prediction models to the PGC-PTSD data (9,245 cases/ 24,285 controls). Models included 12 brain regions, five cardiovascular tissues, 2 endocrine tissues, the tibial nerve, adipose tissue and whole blood. Results We identified 24 significant gene-tissue associations, of which 5 were in peripheral tissues (adrenal gland, heart atrial appendage, tibial artery, tibial nerve). We stratified analyses according to trauma type (civilian vs. combat trauma), sex, and self-defined ancestry. Our three strongest associations were identified in military cohorts only, which supports the hypothesis that there is substantial genetic heterogeneity between civilian and combat PTSD risk. We used the PsychENCODE neuronal and non-neuronal reference map for two histone marks associated with transcription and open chromatin (3-trimethyl-lysine 4 (H3K4me3) and H3-acetyl-lysine 27 (H3K27ac) to understand patterns of histone modification among our PTSD-associated genes. Preliminary analyses indicate a significant correlation between PTSD-association statistics and the presence of both histone marks (correlation with neuronal H3K4me3, Pearson R=0.87, p=3.99×10-5). We intend to expand this analysis to include a wider range of histone modification marks. Discussion We will further expand these analyses to identify tissue specific gene clusters and enriched pathways across tissues or in specific tissues. Finally, we will use neuroimaging data and physiological cardiovascular data to functionally validate our results.