Multivariate, Multi-omic Analysis in 799,429 Individuals Identifies 134 Loci Associated 
with Somatoform Traits

doi:10.21203/rs.3.rs-4823644/v1

Download PDF

Article

Multivariate, Multi-omic Analysis in 799,429 Individuals Identifies 134 Loci Associated with Somatoform Traits

https://doi.org/10.21203/rs.3.rs-4823644/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Somatoform traits, which manifest as persistent physical symptoms without a clear medical cause, are prevalent and pose challenges to clinical practice. Understanding the genetic basis of these disorders could improve diagnostic and therapeutic approaches. With publicly available summary statistics, we conducted a multivariate genome-wide association study (GWAS) and multi-omic analysis of four somatoform traits—fatigue, irritable bowel syndrome, pain intensity, and health satisfaction—in 799,429 individuals genetically similar to Europeans. GWAS identified 134 loci significantly associated with a somatoform common factor, including 44 loci not significant in the input GWAS and 8 novel loci for somatoform traits. Gene-property analyses highlighted enrichment of genes involved in synaptic transmission and enriched gene expression in 12 brain tissues. Six genes, including members of the CD300 family, had putatively causal effects mediated by protein abundance. There was substantial polygenic overlap (76–83%) between the somatoform and externalizing, internalizing, and general psychopathology factors. Somatoform polygenic scores were associated with obesity, Type 2 diabetes, tobacco use disorder, and mood/anxiety disorders in independent biobanks. Drug repurposing analyses suggested potential therapeutic targets, including MEK inhibitors. Mendelian randomization indicated protective effects of gut microbiota, including Ruminococcus bromii. These biological insights provide promising avenues for treatment development.

Biological sciences/Genetics/Genetic association study/Genome-wide association studies

Health sciences/Pathogenesis/Clinical genetics

Biological sciences/Psychology

Persistent physical symptoms (PPS) adversely impact the quality of life of affected individuals, increase healthcare utilization, and contribute to disability as much as or more than etiologically well-defined medical diseases.¹ PPS have historically been referred to as medically unexplained symptoms, but this terminology fails to reflect the dynamic nature of medical knowledge and contributes to a false contrast between these symptoms and “real” medical diseases.² PPS may occur alone or as part of a functional somatic syndrome (FSS), such as irritable bowel syndrome (IBS), fibromyalgia, or myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). It is estimated that up to 10% of the population is affected by at least one FSS,³ with higher rates likely at the symptom level. In a meta-analysis of over 70,000 primary care patients from 24 countries, ~ 45% reported at least one somatic symptom with no identified organic cause.⁴ Symptoms that general practitioners deem medically unexplained occur at higher rates among women and non-native speakers,⁵ potentially reflecting biases in diagnostic practices and healthcare delivery. Despite their prevalence and impact, the underlying mechanisms of PPS remain poorly understood.

The various presentations of PPS have complex genetic and environmental etiologies, further complicated by frequently co-occurring psychiatric disorders.^1,6 Genome-wide association studies (GWAS), which aim to identify genetic risk variants for complex traits and diseases, have revealed loci associated with IBS,⁷ chronic pain,⁸ pain intensity,⁹ and headaches/migraine.¹⁰ These GWAS implicate gene expression and biological processes in the immune and central nervous systems as having key roles in the etiology of PPS. Furthermore, significant genetic correlations exist between different PPS and between PPS and autoimmune, psychiatric, and anthropometric traits.^9,11,12 The substantial phenotypic and genetic correlations between different PPS and between PPS and psychiatric conditions suggest a shared etiology.^13,14 Aligned with this perspective, the Hierarchical Taxonomy of Psychopathology (HiTOP)—an empirical nosology model—proposes that a somatoform spectrum reflects the common liability to PPS.^15,16 Furthermore, reflecting the commonality of somatoform traits with psychiatric conditions, the HiTOP somatoform spectrum is subsumed under a broader emotional dysfunction superspectrum.¹⁶

Previous work using genomic structural equation modeling (gSEM) provided evidence that chronic pain conditions have a shared genetic basis.¹² Other studies demonstrated shared genetic variation among six nociplastic pain conditions (i.e., those in which there is persistent pain without tissue damage): chronic widespread pain (CWP), endometriosis, low back pain, broadly defined headache, irritable bowel syndrome (IBS), and temporomandibular joint disorder (TMJD).¹¹ However, this past work focused largely on pain rather than the broader somatoform spectrum, and some of the included GWAS had modest sample sizes (e.g., TMJD: N_cases = 217, CWP: N_cases = 6,914), which may have limited the strength and precision of the identified common factor. A deeper understanding of the shared genetic basis across a broader spectrum of PPS could identify contributory factors and biological pathways that underpin these syndromes. Genetic research on somatoform traits and their potential polygenic overlap with psychiatric conditions could also help refine the current diagnostic system, improving the assessment and treatment of PPS.

Because our understanding of the genetic basis of the broader PPS spectrum remains limited, we leveraged gSEM to examine the shared genetic architecture of fatigue, health satisfaction, IBS, and pain intensity. Although we initially included headache/migraine as well, it was not retained due to poorer fit on the latent factor. These traits index a possible somatoform spectrum, as they encompass PPS and symptom perceptions.¹⁶ Next, we performed a comprehensive set of bioinformatic analyses using the GWAS summary statistics (Fig. 1). Gene prioritization efforts included gene mapping, functional annotation, transcriptome-wide association studies (TWAS) in enriched tissues, and the integration of brain eQTL and blood plasma pQTL data using summary-data-based Mendelian randomization (SMR). To characterize the genetic architecture of the somatoform factor more broadly, we performed MAGMA gene-property analyses, univariate and bivariate causal mixture models (MiXeR), genetic correlations with 1,426 publicly available GWAS, and phenome-wide association studies (PheWAS) in three cohorts (i.e., BioVU, Penn Medicine BioBank, and Yale-Penn). Finally, we conducted two sets of analyses aimed at translational applications of GWAS—drug repurposing and Mendelian randomization to identify causal effects of the gut microbiome on the somatoform factor given emerging evidence of the role of the gut-brain-axis in regulating pain.¹⁷ With this approach, we aimed to deepen our understanding of the genetic basis of somatoform traits, contribute to the refinement of existing diagnostic models, and identify potential treatments.

Summary Statistics

Summary statistics were chosen to correspond to the HiTOP somatoform factor,¹⁵ which consists of conversion (i.e., neurological symptoms that lack an identified medical cause), somatization (i.e., the tendency to express psychological distress in the form of physical symptoms), malaise (i.e., a general sense of poor wellbeing), head pain, gastrointestinal, and cognitive (e.g., illness anxiety) symptoms. We acknowledge the historical stigma associated with terms like “conversion” and “somatization”¹⁸ and use them here solely to establish consistency with the HiTOP model. In line with patient advocacy groups, we encourage the use of alternate terms like FSS and PPS.

We selected five sets of publicly available GWAS summary statistics from well-powered studies of individuals genetically similar to Europeans (EUR):¹⁹ fatigue (N = 350,580; http://www.nealelab.is/uk-biobank/), headache and migraine¹⁰ (N = 360,391; not retained due to poorer common factor fit), health satisfaction (N = 119,567; http://www.nealelab.is/uk-biobank/), IBS⁷ (N = 486,601), and pain intensity⁹ (N = 436,683). The input summary statistics included between 6,788,440 (pain intensity) and 13,586,245 (fatigue) SNPs. Health satisfaction was reverse coded so that higher scores indicated lower satisfaction to maintain consistency of risk direction with the other traits.

Genomic Structural Equation Modeling

Using the GenomicSEM R package,²⁰ standard quality controls were applied to the input summary statistics, including filtering to EUR HapMap3 SNPs,²¹ and when the information was available, retaining only SNPs with MAF > 0.01 and INFO > 0.6. Linkage disequilibrium score regression (LDSC) was implemented within GenomicSEM to estimate genetic covariance matrices for the input traits. The resulting LDSC matrices were then used to perform confirmatory factor analysis (CFA) to determine whether the hypothesized common factor model fit the data well. Model fit was evaluated based on chi-square (non-significance indicates better fit), Akaike information criterion (AIC; lower values indicate better relative fit), comparative fit index (CFI; > 0.9 indicates good fit), and standardized root mean square residual (SRMR; < 0.08 indicates good fit) values. We also examined the proportion of variance in each input trait that was explained by the common factor to ensure that each trait was sufficiently represented by the factor, aiming for standardized loadings of at least 0.30.

To perform common factor GWAS, we regressed each SNP on the somatoform latent variable. Q_SNP was used to identify any SNPs that had heterogeneous effects across the input traits. SNPs with a significant Q_SNP value (p < 5e-8) were removed from the summary statistics and not included in any subsequent analyses, as these SNPs’ effects were not well represented by a common factor. The effective sample size of the resulting GWAS was calculated using the formula described by Mallard et al. (2022).²² Clumping of the GWAS results was performed using PLINK 1.9²³ with an r² threshold of 0.1 and a physical distance threshold of 3000 kb. The novelty of loci was based on the lead SNP’s positional overlap (within \(\:\pm\:\) 1000 kb) with genome-wide significant (GWS) variants from previous GWAS of the included traits. For all novel SNPs, we performed SNP-level PheWAS using both GWAS Atlas²⁴ and the ieugwasr R package from the OpenGWAS database.²⁵

Gene Mapping and Functional Annotation

Using the SNP2GENE function in FUMA v1.5.2²⁶, SNPs were mapped to protein-coding genes using three approaches: (1) position (\(\:\le\:\) 10kb), (2) eQTL (BrainSeq,²⁷ PsychENCODE,²⁸ BRAINEAC,²⁹ and GTEx v8³⁰ brain tissue data), and (3) chromatin interaction mapping (Hi-C brain tissues).^31,32 SNPs were functionally annotated using ANNOVAR,³³ Combined Annotation Dependent Depletion (CADD),³⁴ RegulomeDB,³⁵ and PsychENCODE²⁸ databases. ANNOVAR identifies the functional consequences of SNPs, such as whether they fall within exons, introns, or regulatory regions. Functional predictions also included protein-coding changes (synonymous, nonsynonymous, stop-gain/loss) and splicing effects. CADD scores were used to assess the deleteriousness of SNPs, with higher scores indicating a greater likelihood of pathogenicity. We considered SNPs with CADD scores above 20 to be potentially deleterious and those above 12.37 (top 7.5%) to be likely pathogenic.³⁶ RegulomeDB v2.2 was used to predict the regulatory potential of SNPs using functional data from eQTL, chromatin states, transcription factor binding sites, and other regulatory elements.³⁷ Lower RegulomeDB scores indicate higher confidence in regulatory function. PsychENCODE data provided additional context on the regulatory role of SNPs in brain tissues.

MAGMA Gene-Based Analyses

We performed gene-set and gene-property analyses using MAGMA v1.08.³⁸ First, SNPs were positionally mapped (0 kb window) to protein-coding genes. Using the resulting gene-level p-values, gene-set analyses were performed for MsigDB v7.0³⁹ curated gene sets and gene ontology (GO) terms. Gene-property analyses were performed for 54 tissue types (GTEx v8)³⁰ and 11 developmental stages in brain samples (BrainSpan).³² These analyses test whether genes having certain properties (i.e., expression in a particular tissue or developmental period) are more likely to be associated with the somatoform factor, adjusting for the average expression of all tissues/developmental stages in the dataset. Additionally, cell-type-specificity analyses were conducted across 16 human brain cell datasets to investigate whether specific cell types were implicated in somatoform traits.

eQTL and pQTL Association Analyses

To identify SNPs with associations to somatoform traits mediated by effects on gene and protein expression, we performed summary-data-based Mendelian randomization (SMR) analyses.⁴⁰ The heterogeneity in dependent instruments (HEIDI) test was used to distinguish pleiotropic effects from linkage between somatoform traits and eQTL/pQTLs. We identified significant associations as those that had a P_SMR < 0.05 after Bonferroni correction for the number of genes tested and a P_HEIDI > 0.05. We used the MetaBrain⁴¹ cis-eQTL database from 7 brain regions due to its large sample size (n = 8,613) and comprehensive integration of data across 14 cohorts. For the blood pQTL analyses, we used two cis- and trans-pQTL databases. The first was the UK Biobank Pharma Proteomics Project (UK-PPP) that included samples from 54,219 individuals and measured 2,923 unique proteins.⁴² The second database was from the deCODE Consortium, including samples from 35,559 individuals and 4,719 unique proteins.⁴³ Use of the two databases allowed for validation and replication of findings and leveraged varied proteomic approaches for more comprehensive protein coverage.

Transcriptome-Wide Association Studies

We conducted two transcriptome-wide association studies (TWAS) using MetaXcan software.^44,45 In the first, we used S-MultiXcan to simultaneously examine associations across all GTEx v8 tissues for which gene expression was enriched based on MAGMA results. S-MultiXcan accounts for the correlation between gene expression profiles across different tissues, increasing statistical power to detect biologically meaningful associations.⁴⁵ We complemented this approach with S-PrediXcan analyses using data from PsychENCODE (i.e., 2,188 postmortem frontal and temporal cerebral cortex samples from 1,695 adults),⁴⁶ which is enriched for individuals diagnosed with psychiatric conditions that are often comorbid with functional somatic conditions.⁶ This provides a more clinically relevant cohort for studying transcriptomic associations.

Polygenic Overlap with Psychopathology

MiXeR software^53,54 was used to conduct univariate and bivariate causal mixture models. Univariate causal mixture models estimate the polygenicity (i.e., the number of causal variants needed to explain 90% SNP-heritability) and discoverability (i.e., the average effect size of causal variants) of traits. Bivariate causal mixture models can estimate genetic overlap between traits, even when the causal variants have opposite directions of effect on the traits. The Dice coefficient estimates the proportion of polygenic SNPs out of the total number of estimated causal SNPs for both traits. We performed bivariate MiXeR analyses to estimate the degree of polygenic overlap between the somatoform factor, externalizing, internalizing, and general psychopathology factors.⁵⁵ We selected these factors because of their inclusion in the HiTOP model⁵⁶ and the high rates of comorbidity between somatoform traits and psychopathology.³

Genetic Correlations with Publicly Available GWAS

We performed batch genetic correlations with 1,426 phenotypes from publicly available GWAS using the Complex-Trait Genetics Virtual Lab (CTG-VL).⁴⁷ The included phenotypes spanned a variety of domains, including biological variables, physical diseases, and psychiatric disorders. CTG-VL uses LDSC software^48,49 with 1000 Genomes Project phase 3⁵⁰ EUR data as LD references. We applied a Bonferroni correction (p = 0.05/1426 = 3.51e-5) to account for multiple testing and identify significant correlations.

Lab- and Phenome-Wide Association Studies

We conducted phenome-wide association studies (PheWAS) using somatoform polygenic scores (PGS) in three cohorts: BioVU, Penn Medicine BioBank (PMBB), and Yale-Penn. ICD-9 and ICD-10 codes from electronic health records (EHRs) or diagnostic interviews were mapped to phecodes. Lab-wide association studies (LabWAS) were also performed in BioVU to examine associations with lab test results and biomarkers.⁵¹ We calculated PGS using PRS-CS software,⁵² applying the default settings to estimate shrinkage parameters. PheWAS analyses were conducted in the PheWAS v0.12 R package using logistic or linear regression models, depending on the phenotype. Analyses were restricted to phenotypes that had at least 100 cases (for binary traits) or individuals assessed (for continuous traits). All models included age, sex, and the first ten genetic ancestry principal components (PCs) as covariates, with a Bonferroni correction applied to identify significant associations. Given that both BioVU and PMBB are EHR datasets, we meta-analyzed the PheWAS results from the two cohorts. For each unique phenotype (1,442 total), we pooled the effect sizes by calculating a weighted average using the beta estimates and standard errors from each cohort.

BioVU. The BioVU cohort comprises Vanderbilt University Medical Center patients with EHR and genotype data.⁵³ As described elsewhere,⁵¹ genotyping was performed using the Illumina Multi-Ethnic Genotype Array (MEGAEX). Genotypes were filtered for SNP (< 0.95) and individual (< 0.98) call rates, sex discrepancies, and excessive heterozygosity (|F_het| > 0.2).⁵⁴ PCA was performed using FlashPCA2 1000 Genomes phase 3 reference datasets⁵⁰ to identify European-like individuals. Genotypes were imputed using the Michigan Imputation Server (https://imputationserver.sph.umich.edu) with the Haplotype Reference Consortium panel.⁵⁵ SNPs with imputation quality R2 > 0.3 or INFO > 0.95 and MAF > 0.01 were retained. PheWAS analyses were performed in the EUR cohort (up to 66,214 individuals) on 1,338 case/control phecodes. To identify associations with biomarkers, we performed LabWAS for 315 biomarkers using a pipeline developed by Dennis et al.⁵¹

Penn Medicine BioBank. The Penn Medicine BioBank (PMBB) comprises a cohort recruited through the University of Pennsylvania Health System.⁵⁶ Genotyping was performed using the GSA array, phasing was performed using EAGLE,⁵⁷ and imputation was performed using Minimac4⁵⁸ on the TOPMed Imputation server.⁵⁸ Variants with imputation quality < 0.7, missingness > 5%, MAF < 1%, and sample call rate < 0.99 were excluded. PCs for genetic ancestry were calculated in EIGENSOFT v7.2.0 (https://github.com/DReichLab/EIG). Genetically inferred ancestry was assigned based on the distance of ten PCs from the 1000 Genomes⁵⁰ reference populations. We performed PheWAS using 920 phenotypes in up to 29,090 EUR individuals.

Yale-Penn. The Yale-Penn sample, enriched for individuals with substance use disorders (SUDs), includes 5,424 EUR individuals with genotype data.⁵⁹ As described previously,⁵⁹ genotyping was performed using the Illumina HumanOmni1-Quad microarray, the Illumina HumanCoreExome array, or the Illumina Multi-Ethnic Global array. Imputation was performed using Minimac3⁵⁸ and the 1000 Genomes Project phase 3⁵⁰ reference panel on the Michigan Imputation Server (https://imputationserver.sph.umich.edu). SNPs with imputation quality < 0.7, MAF < 0.01, missingness > 0.01, or a batch allele frequency difference > 0.04 were excluded, as were individuals with genotype call rate < 0.95.⁵⁹ PCs were used to determine genetic similarity to 1000 Genomes Project phase 3⁵⁰ reference genomes. Yale-Penn participants were interviewed with the Semi-Structured Assessment for Drug Dependence and Alcoholism (SSADDA).⁶⁰ PheWAS analyses were performed for 622 traits.

Drug Repurposing

To perform drug repurposing, we used the Library of Integrated Network-Based Cellular Signatures (LINCS) L1000 database, which catalogs in vitro gene expression profiles from thousands of chemical compounds across more than 80 human cell lines. We focused on compounds that are either approved by the Food and Drug Administration (FDA) or currently undergoing clinical trials (https://clue.io/repurposing#download-data),⁶¹ yielding a total of 829 compounds (590 of which are FDA-approved) and expression profiles from five neuronal cell lines. This resulted in 3,897 unique signatures. We then matched medication signatures to gene expression signatures obtained from a somatoform TWAS performed on 12 enriched brain tissues. Weighted Pearson correlations were calculated between each brain transcriptome association and the compound signatures,⁶² with genes weighted by the proportion of heritability explained using the metafor R package (v.3.8-1). Each compound was included as a fixed effect, incorporating the weighted effect size (r_weighted) and sampling variability (se2_r_weighted) from all its signatures across different times, cell lines, and doses. To account for potential heterogeneity across tissues, brain region was included as a random effect. Significance was evaluated using a Bonferroni correction (p = 0.05/829 = 6.03e-5).

We also used a second genetically informed drug repurposing method, drug-gene set analysis (DRUGSETS),⁶³ with data sourced from the Drug Repurposing Hub⁶¹ and the Drug Gene Interaction Database.⁶⁴ For this method, drug–gene sets were created for 1,201 drugs, consisting of genes whose protein products are targeted by or interact with each drug. Competitive gene-set analysis was conducted using MAGMA v.1.08,³⁸ conditioning on a comprehensive set of all drug targeted genes in the data (n = 2,116 across 735 gene sets). To determine significant drug-gene sets, we applied a Bonferroni correction (p = 0.05/735 = 6.80e-5).

Mendelian Randomization with Gut Microbiome Taxa

To identify potential causal effects of the gut microbiome on somatoform traits, we conducted Mendelian randomization (MR) analyses using TwoSampleMR version 0.5.10 in R.⁶⁵ We selected two GWAS of gut microbiota abundance to examine in relation to somatoform traits based on their complementary strengths. The first included 211 taxa (with genus being the most granular level) measured in 14,306 EUR individuals.⁶⁶ The second included 207 taxa, including species level information, measured in a smaller sample of 7,738 Dutch individuals.⁶⁷ Through the IEU GWAS database, we extracted instruments for each of the taxa at a p-value threshold of 1e-5. Instruments were clumped using EUR 1000 Genomes Project data to ensure independence. Matching SNPs were then extracted from the somatoform GWAS for use as outcomes. If an instrument SNP was not available in the somatoform GWAS, we selected proxy SNPs using the default settings in TwoSampleMR. The inverse variance weighted estimate was the primary MR method used for inferring causal effects. As sensitivity analyses, we performed tests of heterogeneity and horizontal pleiotropy, which assess whether the causal effects are consistent across all instruments and whether the instruments affect the outcome through pathways other than the exposure. We also performed the Steiger test, which can be used to evaluate whether the specified causal direction (i.e., microbiome → somatoform) is likely to be correct or if effects may operate in the opposite direction.⁶⁸

LDSC identified significant genetic correlations between all the included traits (Supplementary Fig. 1). However, genetic correlations with headache/migraine were lower than the other traits (rg_average = 0.27 vs. 0.58). A common factor CFA fit well (\(\:\chi\:\)²(5) = 21.61, p = 0.001, AIC = 41.61, CFI = 0.99, SRMR = 0.03; Fig. 2b). Standardized loadings ranged from 0.35 (for headache/migraine) to 0.89 (for fatigue). Although the loading of headache/migraine met our pre-specified minimum of 0.30, it was substantially lower than the other traits (all > 0.65), suggesting that headache/migraine was not adequately represented by the common factor. To maintain a higher standard of representation and consistency, we excluded headache/migraine from the final model. After its exclusion, model fit improved significantly (\(\:\varDelta\:\chi\:\)²(3) = 12.31, p = 0.006; fit statistics: \(\:\chi\:\)²(2) = 9.30, p = 0.01, AIC = 25.30, CFI = 1, SRMR = 0.03).

GWAS of the somatoform factor identified 134 significantly associated loci (N_eff = 799,429; Supplementary Table 1), of which 44 were not GWS in any of the input GWAS, and 8 had not been previously associated with somatoform traits (Fig. 2a). The most strongly associated locus (p = 6.80e-16) was outside the MHC region of chromosome 6 nearest to gene UHRF1BP1, which is involved in assisting or regulating epigenetic modifications.⁶⁹ The lead SNP in this locus (rs9469907) functions as an eQTL for several genes, including C6orf106, SNRPC, and CLPS, in PsychENCODE and GTEx brain tissues.

In SNP-level PheWAS (Supplementary Table 2 and Supplementary Fig. 2), the lead SNPs from the 8 novel loci were associated primarily with physical health, immunological, and mental health measures (e.g., insomnia, C-reactive protein levels, body mass index, wellbeing, depression, and Type 2 diabetes). Based on the Q_SNP test, 8 loci exhibited heterogeneous effects across the somatoform traits (Supplementary Table 3). Of the lead SNPs from these loci, all but one showed the strongest association with pain intensity. The exception, rs10795616, was most strongly associated with health satisfaction (p = 2.21e-08).

Gene Mapping and Functional Annotation

MAGMA identified 874 genes based on position, eQTL, and chromatin interactions. One-third (33.98%; n = 297) of genes were mapped by more than one approach, and 11.21% (n = 98) were mapped by all three. eQTL and chromatin interaction plots are in Supplementary Fig. 3. Candidate SNPs showed an enrichment of intronic (p = 7.95e-317), intergenic (p = 7.43e-297), non-coding RNA intronic (p = 1.08e-8), 3' UTR (p = 7.87e-7), and 5' UTR (p = 2.62e-4) functional categories (Fig. 2c). Of candidate SNPs (n = 5,555), 30 had a CADD score > 20, and 255 (4.59%) had scores ≥ 12.37, which is suggestive of deleteriousness to gene function (Supplementary Table 4). Of the candidate SNPs, 59.17% (n = 3287) had RegulomeDB scores indicative of regulatory functions related to transcription factor binding and gene expression (i.e., 1a-1f; Supplementary Table 5).

Gene-Based Enrichment

MAGMA gene-set analyses showed enrichment for genes involved in negative regulation of synaptic transmission (b = 0.71, SE = 0.14, p = 3.24e-07). Gene-property analyses identified significant gene enrichment in 11 GTEx v8 brain tissues (Fig. 3a). The strongest associations were for the cerebellar hemisphere (b = 0.05, SE = 0.01, p = 1.43e-12), cerebellum (b = 0.05, SE = 0.01, p = 2.63e-12), and frontal cortex (b = 0.05, SE = 0.01, p = 1.97e-9). In addition, there was significant enrichment for gene expression in the pituitary gland (b = 0.05, SE = 0.01, p = 5.64e-6). Gene-property analyses using BrainSpan brain samples from 11 developmental stages showed enrichment for gene expression in the early mid-prenatal (b = 0.04, SE = 0.01, p = 0.001) and late mid-prenatal (b = 0.05, SE = 0.02, p = 0.002) stages (Supplementary Fig. 4).

To investigate cell-specific enrichment, we performed cell-type specificity analyses using 16 human brain scRNA-seq datasets and identified 7 cell-specific gene expression profiles associated with the somatoform factor after multiple testing correction (Fig. 3b). Of the seven, three (GABAergic neurons in the prefrontal cortex at gestational week 26, GABAergic neurons in the human midbrain, and inhibitory neurons from PsychENCODE adult brain samples) were independently associated with the somatoform factor, with the others jointly explained by their association with the independent cell types. In cross-dataset conditional analyses, GABAergic neurons at gestational week 26 and inhibitory adult neurons were significantly collinear, suggesting that the associations of these two cell types are driven by similar genetic signals.

SMR with Brain eQTLs and Blood Plasma pQTLs

We conducted summary-data-based Mendelian randomization (SMR) analyses to examine whether the associations of SNPs with somatoform traits were mediated by effects on gene expression in the brain and protein expression in blood plasma. We identified 28 genes whose expression levels in the brain exerted putatively causal effects on somatoform traits (Bonferroni-adjusted p < 0.05 and p_SMR > 0.05; Fig. 4a). Among these were UHRF1BP1, which is known to interact with a key regulator of DNA methylation, and HLA-DRB1, part of the human leukocyte antigen (HLA) family of genes that is critical in initiating immune responses and implicated in many autoimmune conditions.

Using data from the UKB Pharma Proteomics Project (UKB-PPP), we identified 117 genes whose protein levels were significantly associated with somatoform traits, including several members of the CD300 family that are involved in modulating inflammatory responses (Fig. 4b). Additionally, genes involved in neural development and synaptic processes were significant, such as HS6ST1 and LRRN1. Using plasma proteomic data from deCODE, we identified 57 genes whose effects on protein abundance were putatively causally associated with somatoform traits (Fig. 4c). Across the UKB-PPP and deCODE analyses, 6 genes were consistently identified: two members of the CD300 family of genes (i.e., CD300A and CD300C), CLEC4G, HS6ST1, LRRN1, and RNASET2.

Transcriptome-Wide Association Analyses

To examine genes whose expression in enriched tissues was related to somatoform traits, we performed two TWAS using MetaXcan.^44,45 In the first, we used S-MultiXcan to simultaneously examine gene expression profiles across the 12 tissues that showed significant enrichment in MAGMA analyses. After Bonferroni correction (p = 0.05/14,368 = 3.48e-6), we identified 158 genes with significantly altered expression levels associated with somatoform traits (Fig. 5a). In a second TWAS of brain tissues in psychiatric cases and controls,⁷⁰ we identified 131 genes that showed significantly different (p \(\:\le\:\) 4.16e-6) expression levels (Fig. 5b). Across the two TWAS, 34 genes were consistently identified, including genes related to immune function (e.g., MST1R), oxidative stress response (e.g., GPX1), neural development (e.g., DPYSL5 and SORCS3), and neurotransmitter signaling (e.g., GRK4). Several showed enriched expression across brain tissues (e.g., DPYSL5 and GPX1; Supplementary Fig. 5). Across the two TWAS and the SMR analysis, five genes (CCDC144CP, PPP6C, SCAI, UHRF1BP1, USP32P3) were consistently implicated.

Polygenicity, Discoverability, and Polygenic Overlap with Psychopathology

We used MiXeR software to conduct univariate and bivariate causal mixture models to assess the polygenicity (i.e., the number of variants estimated to be needed to explain 90% SNP-heritability) and discoverability (i.e., the causal effect size variance) of somatoform traits, as well as their polygenic overlap with psychopathology spectra (Supplementary Fig. 6). MiXeR models estimated 11,321 causal SNPs for the somatoform factor (SD = 562.64), with an average discoverability of 7.50e-6 (SD = 3.55e-7). The somatoform factor and externalizing psychopathology were moderately genetically correlated (r_g = 0.46, SD = 0.02), and shared an estimated 76% of their causal variants (SD = 0.18). Similarly, the somatoform factor and internalizing psychopathology were significantly genetically correlated (r_g =0.63, SD = 0.01) and shared 79% of their causal variants (SD = 0.12). Estimates were also similar for the polygenic overlap between general psychopathology and the somatoform factor (r_g = 0.69, SD = 0.01, Dice = 0.83, SD = 0.07).

Genetic Correlations

Genetic correlations were performed between the somatoform factor and 1,426 publicly available GWAS. After Bonferroni correction (p = 0.05/1,426 = 3.51e-5), the somatoform factor was significantly genetically correlated with 646 phenotypes (Supplementary Table 6 and Supplementary Fig. 7). Consistent with the somatoform factor representing PPS for which there is no identifiable medical cause, one of the top genetic correlations was with “symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified” (r_g = 0.80, SE = 0.02, p = 1.69e-242). Of the significant associations, at least 70 (10.84%) were with pain-related conditions and medications. Many psychiatric traits were also significantly genetically correlated with the somatoform factor, including major depressive disorder (r_g = 0.60, SE = 0.02, p = 3.29e-212), mood swings (r_g = 0.60, SE = 0.02, p = 1.73e-166), and loneliness/isolation (r_g = 0.62, SE = 0.02, p = 6.02e-136). Results identified significant genetic correlations with obesity-related phenotypes (e.g., waist circumference, body mass index, and whole-body fat mass), socioeconomic status (e.g., educational attainment, unemployment, and financial difficulties), and general health (e.g., having a longstanding illness, taking prescription medications, and lacking physical activity).

Lab- and Phenome-Wide Association Scans

BioVU and PMBB Meta-Analysis. After meta-analyzing effects across BioVU and PMBB, there were 229 significant associations with the somatoform PGS (Fig. 6). The top associations included obesity (beta = 0.18, SE = 0.01, p = 2.76e-76), tobacco use disorder (TUD; beta = 0.18, SE = 0.01, p = 9.69e-72), type 2 diabetes (beta = 0.17, SE = 0.01, p = 3.67e-69), and mood disorders (beta = 0.14, SE = 0.01, p = 3.00e-52). There were many significant associations with pain-related conditions, including nonspecific chest pain (beta = 0.12, SE = 0.01, p = 2.36e-38), unspecified muscle pain (beta = 0.18, SE = 0.02 p = 3.64e-29), abdominal pain (beta = 0.09, SE = 0.01, p = 2.18e-27), and chronic pain (beta = 0.12, SE = 0.01, p = 2.83e-22). Study-specific results from BioVU and PMBB are in Supplementary Tables 7 and 8.

BioVU LabWAS. After Bonferroni correction, LabWAS identified 40 significant associations of the somatoform factor with biomarkers (Supplementary Fig. 8), including elevated C reactive protein levels (CRP; beta = 0.03, SE = 0.005, p = 4.43e-15) and erythrocyte sedimentation rates (beta = 0.06, SE = 0.01, p = 1.92e-06). Other notable associations were with higher erythrocyte distribution widths (beta = 0.06, SE = 0.004, p = 6.66e-52) and white blood cell counts (beta = 0.05, SE = 0.004, p = 1.58e-37), and with lower levels of iron (beta = -0.04, SE = 0.01, p = 1.75e-6) and Vitamin D (beta = -0.06, SE = 0.01, p = 5e-19). Consistent with the link with diabetes in the PheWAS, the somatoform PGS was also related to higher hemoglobin A1c levels (beta = 0.03, SE = 0.005, p = 4.43e-15).

Yale-Penn. In the Yale-Penn sample, there were 229 significant PheWAS associations after Bonferroni correction (Fig. 6 and Supplementary Table 9). The sample is enriched for substance use disorders, and these were the most prevalent associations identified. However, other phenotypes like lower educational levels (beta = -0.17, SE = 0.01, p = 1.15e-38), poorer self-reported health rating (beta = -0.18, SE = 0.02, p = 1.29e-25), lower household income (beta = -0.32, SE = 0.04, p = 2.19e-17), and greater childhood adversity (beta = 0.28, SE = 0.04, p = 1.16e-12) were also significant in the deeply phenotyped sample.

Drug Repurposing

We used LINCS to match medication signatures to somatoform gene expression signatures. After Bonferroni correction, we identified 324 perturbagens (Supplementary Table 10) comprising a wide array of mechanisms and classes across both emerging (i.e., undergoing clinical trials; n = 113) and FDA-approved (n = 211) therapeutics. Several were of relevance to PPS, including analgesics (e.g., diclofenac, ibuprofen, and venlafaxine), antidiarrheals (e.g., loperamide), and antidepressants (e.g., bupropion). Of note, two identified drugs targeted the MAP2K1 gene and reversed the gene expression signature found in the PsychENCODE TWAS. Both drugs (pd-0325901 and selumetinib) are kinase inhibitors, with selumetinib having been approved to treat symptomatic plexiform neurofibromas in children with a genetic disorder that causes tumors to grow along nerves.⁷¹ Additionally, of the identified perturbagens, ten had gene targets that mapped to GWS SNPs (Supplementary Fig. 9). These included four dopamine receptor antagonists (i.e., nemonapride, melperone, benperidol, and carmoxirole) four kinase inhibitors (i.e., vemurafenib, dabrafenib, pf-562271, and sorafenib), an HDM2 antagonist (i.e., serdemetan), and a calcium channel blocker (i.e., nimodipine). After Bonferroni correction, the DRUGSETS analysis identified one significant association with Anatomical Therapeutic Chemical (ATC) code G04B (t = 3.89, p = 5.44e-5), which comprises urological drugs.

Potentially Causal Effects of the Gut Microbiome

Of the 418 taxa in the MiBioGen and Dutch Microbiome Projects, 326 (130 from MiBioGen and 196 from the Dutch Microbiome Project) had genetic variants associated with their abundance at p < 1e-5 and were included in MR analyses. Of these, 23 (10 from MiBioGen and 13 from the Dutch Microbiome Project) exhibited putatively causal effects on the somatoform factor prior to multiple testing correction, but only four remained significant after Benjamini-Hochberg false discovery rate (FDR) correction (all from the Dutch Microbiome Project; Supplementary Fig. 10). One significant result was based on a single SNP so was not interpreted. At the species level, Adlercreutzia equolifaciens (beta = -0.04, SE = 0.01, p = 5.13e-5) and Ruminococcus bromii (beta = -0.05, SE = 0.01, p = 6.65e-4) exhibited putatively causal protective effects on the somatoform factor. Similarly, the genus Adlercreutzia (beta = -0.04, SE = 0.01, p = 5.14e-5) was also significant and potentially protective. In the MiBioGen cohort, the genus Adlercreutzia was not putatively causally associated with the somatoform factor. However, analyses in that cohort showed a consistent direction of effect. In the MiBioGen cohort, the putative effects of Ruminococcus on the somatoform factor were in a matching direction as well, and the association with the genus was significant prior to FDR correction (beta = -0.01, SE = 0.006, p = 0.04). Therefore, findings in MiBioGen provide relatively modest replication of our results.

Steiger directionality tests supported the proposed causal direction of microbiota on the somatoform factor (all ps < 2.05e-5). We used the Egger intercept to evaluate confounding by pleiotropy, but due insufficient instruments, we were only able to evaluate this for Ruminococcus bromii (Egger intercept = -0.02, SE = 0.05, p = 0.76). There was no evidence for heterogeneity in the effects of instruments for Adlercreutzia equolifaciens (Q(1) = 0.02, p = 0.88), Ruminococcus bromii (Q(2) = 3.65, p = 0.16), or Adlercreutzia (Q(1) = 0.03, p = 0.86).

Although often considered distinct in clinical practice, our findings support a common latent genetic factor that contributes to vulnerability to multiple PPS, including fatigue, IBS, and pain intensity, as well as symptom perceptions, including health satisfaction and pain intensity. This aligns with the emerging view of PPS as part of a broader spectrum linked by shared etiological pathways.⁷² Amid calls for improved classification and treatment of PPS and growing recognition of the complex interactions between mind and body,^72,73 our findings provide insights into the shared genetic architecture underlying these traits, opening new avenues for improved clinical care.

By incorporating various somatoform traits in an effective sample size of 799,429 EUR individuals, we identified 134 loci associated with the common genetic liability to somatoform traits, including 8 lead SNPs in loci not previously GWS in relation to any individual PPS. The power of the multivariate approach also allowed us to replicate findings for 44 loci that were not GWS in the input GWAS but had previously been associated with PPS. Functional annotation indicated enrichment of candidate SNPs in regulatory regions, with the majority (59.17%) having functions related to binding and gene expression. This suggests a significant role for gene regulatory mechanisms in the etiology of PPS.

The expression of genes related to somatoform traits was enriched across brain tissues, particularly the cerebellar hemisphere and cerebellum. Recently, the cerebellum has drawn attention for its potential role in modulating the emotional and cognitive elements of pain through its communication with subcortical and cortical regions.⁷⁴ Individuals with chronic pain show altered activation patterns in the cerebellum during pain,⁷⁵ and differences in functional connectivity between the cerebellum and other brain regions are correlated with ratings of pain intensity.⁷⁶ Furthermore, the cerebellum may be associated with salience processing,^74,77 which involves the integration of internal and external sensory information and has been implicated in both chronic pain^77–79 and IBS.⁸⁰ Cell-type specific analyses implicated the role of inhibitory GABAergic neurons in the human midbrain and prefrontal cortex. Mouse models have underscored the crucial role of GABAergic interneurons in regulating sensory sensitivity,⁸¹ and a GWAS of pain intensity also implicated GABAergic neurons.^9,81 Dysregulation of inhibitory control via GABAergic neurons during development may contribute to heightened sensitivity to PPS.

Performing TWAS and SMR analyses, we sought to identify potential causal genes in enriched tissues using brain transcriptomic data. Across the two TWAS and the SMR analysis, five genes (CCDC144CP, PPP6C, SCAI, UHRF1BP1, USP32P3) were consistently implicated, providing strong evidence for their role in somatoform traits. Two of the identified genes (CCDC144CP and USP32P3) are pseudogenes, which have traditionally been considered non-functional but have recently been found to have important functional and regulatory roles that warrant further investigation.⁸² The remaining three genes are involved in cell signaling (PPP6C), cell migration and tumor suppression (SCAI), and regulation of DNA methylation (UHRF1BP1). We also identified six genes whose effects on protein levels mediated their association with somatoform traits, including four involved in immune regulation (CD300A, CD300C, CLEC4G, and RNASET2) and two in neuronal development (HS6ST1 and LRRN1). Preclinical research is needed to examine the mechanisms by which these prioritized genes influence PPS.

We found substantial polygenic associations between the somatoform and internalizing, externalizing, and general psychopathology factors. Genetic correlations with 1,426 publicly available GWAS further highlighted the shared etiology of somatoform traits and psychopathology, with the genetic correlation with depression being among the strongest. This shared genetic architecture may help explain the high comorbidity between PPS and psychopathology, although our analyses do not provide information on causality or the direction of effect.³ In three biobanks, we found that somatoform PGS were associated with numerous physical and mental health conditions, including diabetes, TUD, mood disorders, obesity, post-traumatic stress disorder, and sleep disorders. Collectively, these findings underscore the need for treatment approaches that recognize the interconnectedness of physical and mental health.

Addressing shared risk factors for various PPS could potentially improve outcomes. Thus, we examined potential causal effects of gut microbiota on the somatoform factor and performed drug repurposing to identify treatments relevant to a broad range of PPS. Across two datasets of host-gut microbiome associations, Ruminococcus bromii had the strongest support for potential protective effects on somatoform traits. R. bromii are a keystone bacterial species for their ability to metabolize resistant starch.⁸³ They also contribute to butyrate production in the colon, which is used by beneficial gut microbes.⁸³ A reduced abundance of R. bromii has been associated with chronic pancreatitis⁸⁴ and Crohn’s disease.⁸⁵ Importantly, even short-term dietary changes substantially alter R. bromii abundance in the gut,⁸⁶ suggesting that it is a modifiable target. We also identified perturbagens that may have promise for treating multiple PPS, including 10 compounds targeting genes mapped by GWS variants. Notably, four were dopamine receptor antagonists, including atypical antipsychotics (nemonapride, melperone, and benperidol) and a peripherally active D2 receptor antagonist with antihypertensive properties (carmoxirole).⁸⁷ Two compounds (PD-0325901 and selumetinib) were reversed the brain transcriptomic expression signature of the somatoform factor by targeting the MAP2K1 gene. One of these, PD-0325901, has potent anti-inflammatory activity.⁸⁸ Whereas mitogen-activated protein kinase (MAPK) signaling pathways are involved in pain sensitization and their inhibition reduces pain in animal models,⁸⁹ our findings suggest that there may be broader applications of these inhibitors in treating other PPS as well.

This study has several limitations. First, the summary statistics included were exclusively from EUR individuals, potentially limiting generalizability to individuals genetically similar to non-EUR populations. Future research should aim to replicate findings in other groups as biobanks continue to grow and become more diverse. For example, a study applying gSEM in AFR individuals showed that when GWAS are available and appropriate steps are taken to model the complex LD patterns, these models yield accurate and meaningful results.⁹⁰ Additionally, although genetic correlation and PheWAS analyses underscored the complexity of interactions between physical and mental health, they do not provide information on the mechanisms underlying these associations, which could include environmental risk factors, indirect or direct genetic effects, or other possibilities. Third, associations between gut microbiota and somatoform traits require replication and experimental research to uncover their potential mechanisms. Finally, further research is necessary to validate the efficacy and safety of therapeutic targets identified through drug repurposing analyses, emphasizing the ongoing challenge of translating genetic discoveries into effective treatments.

By identifying a common genetic factor that contributes to vulnerability to multiple PPS, our findings highlight shared biological pathways that link conditions often considered distinct in clinical practice. Enrichment analyses point to the cerebellum and GABAergic neurons as key players in the neurobiology of various PPS. The significant genetic correlations and high polygenic overlap with psychopathology align with dimensional taxonomic approaches. These findings enhance our understanding of PPS, open new avenues for research and therapeutic development, and underscore the need for approaches that address the complex interplay between genetic, neurological, and psychological factors in health and disease.

Acknowledgments: Veterans Integrated Service Network 4, Mental Illness Research, Education and Clinical Center, National Institute on Alcohol Abuse and Alcoholism grants R01 AA030056 and K01 AA028292 (to RLK), and National Human Genome Research Institute grant T32 HG009495 (to KLF).

Conflicts of Interest: Dr. Kranzler is a member of advisory boards for Dicerna Pharmaceuticals, Sophrosyne Pharmaceuticals, Enthion Pharmaceuticals, Clearmind Medicine, and Altimmune; a consultant to Sobrera Pharmaceuticals; the recipient of research funding and medication supplies for an investigator-initiated study from Alkermes; and a member of the American Society of Clinical Psychopharmacology’s Alcohol Clinical Trials Initiative, which was supported in the last three years by Alkermes, Dicerna, Ethypharm, Lundbeck, Mitsubishi, Otsuka, and Pear Therapeutics; Drs. Gelernter and Kranzler hold U.S. patent 10,900,082 titled: "Genotype-guided dosing of opioid agonists," issued 26 January 2021.

Joustra, M.L., Janssens, K.A.M., Bültmann, U. & Rosmalen, J.G.M. Functional limitations in functional somatic syndromes and well-defined medical diseases. Results from the general population cohort LifeLines. Journal of Psychosomatic Research 79, 94–99 (2015).
Marks, E.M. & Hunter, M.S. Medically Unexplained Symptoms: An acceptable term? British Journal of Pain 9, 109–114 (2014).
Petersen, M.W., et al. Irritable bowel, chronic widespread pain, chronic fatigue and related syndromes are prevalent and highly overlapping in the general population: DanFunD. Scientific Reports 10, 3273 (2020).
Haller, H., Cramer, H., Lauche, R. & Dobos, G. Somatoform disorders and medically unexplained symptoms in primary care. Deutsches Ärzteblatt International 112, 279–287 (2015).
Steinbrecher, N., Koerber, S., Frieser, D. & Hiller, W. The prevalence of medically unexplained symptoms in primary care. Psychosomatics 52, 263–271 (2011).
Creed, F. Psychiatric disorders comorbid with general medical illnesses and functional somatic disorders: The Lifelines cohort study. PLOS ONE 18, e0286410 (2023).
Eijsbouts, C., et al. Genome-wide analysis of 53,400 people with irritable bowel syndrome highlights shared genetic pathways with mood and anxiety disorders. Nature Genetics 53, 1543–1552 (2021).
Johnston, K.J.A., et al. Genome-wide association study of multisite chronic pain in UK Biobank. PLOS Genetics 15, e1008164 (2019).
Toikumo, S., et al. A multi-ancestry genetic study of pain intensity in 598,339 veterans. Nature Medicine 30, 1075–1084 (2024).
Meng, W., et al. A meta-analysis of the genome-wide association studies on two genetically correlated phenotypes suggests four new risk loci for headaches. Phenomics 3, 64–76 (2023).
Johnston, K.J.A., Signer, R. & Huckins, L.M. Chronic overlapping pain conditions and nociplastic pain. medRxiv, 2023.2006.2027.23291959 (2024).
Zorina-Lichtenwalter, K., et al. Genetic risk shared across 24 chronic pain conditions: identification and characterization with genomic structural equation modeling. Pain 164, 2239–2252 (2023).
Nimnuan, C., Rabe-Hesketh, S., Wessely, S. & Hotopf, M. How many functional somatic syndromes? Journal of Psychosomatic Research 51, 549–557 (2001).
Wessely, S., Nimnuan, C. & Sharpe, M. Functional somatic syndromes: One or many? The Lancet 354, 936–939 (1999).
Kotov, R., et al. The Hierarchical Taxonomy of Psychopathology (HiTOP): A dimensional alternative to traditional nosologies. Journal of Abnormal Psychology 126, 454–477 (2017).
Watson, D., et al. Validity and utility of Hierarchical Taxonomy of Psychopathology (HiTOP): III. Emotional dysfunction superspectrum. World Psychiatry 21, 26–54 (2022).
Guo, R., Chen, L.-H., Xing, C. & Liu, T. Pain regulation by gut microbiota: Molecular mechanisms and therapeutic potential. British Journal of Anaesthesia 123, 637–654 (2019).
Rommelfanger, K.S., et al. Disentangling stigma from functional neurological disorders: Conference report and roadmap for the future. Frontiers in Neurology 8(2017).
National Academies of Sciences, E. & Medicine. Using Population Descriptors in Genetics and Genomics Research: A New Framework for an Evolving Field, (The National Academies Press, Washington, DC, 2023).
Grotzinger, A.D., et al. Genomic structural equation modelling provides insights into the multivariate genetic architecture of complex traits. Nature Human Behaviour 3, 513–525 (2019).
The International HapMap 3 Consortium. Integrating common and rare genetic variation in diverse human populations. Nature 467, 52–58 (2010).
Mallard, T.T., et al. Multivariate GWAS of psychiatric disorders and their cardinal symptoms reveal two dimensions of cross-cutting genetic liabilities. Cell Genomics 2(2022).
Chang, C.C., et al. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience 4, 7 (2015).
Watanabe, K., et al. A global overview of pleiotropy and genetic architecture in complex traits. Nature Genetics 51, 1339–1348 (2019).
Elsworth, B., et al. The MRC IEU OpenGWAS data infrastructure. bioRxiv, 2020.2008.2010.244293 (2020).
Watanabe, K., Taskesen, E., van Bochoven, A. & Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. Nature Communications 8, 1826 (2017).
Schubert, Christian R., et al. BrainSeq: Neurogenomics to drive novel target discovery for neuropsychiatric disorders. Neuron 88, 1078–1083 (2015).
Wang, D., et al. Comprehensive functional genomic resource and integrative model for the human brain. Science 362, eaat8464 (2018).
Ramasamy, A., et al. Genetic variability in the regulation of gene expression in ten regions of the human brain. Nature Neuroscience 17, 1418–1428 (2014).
The GTEx Consortium, et al. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318–1330 (2020).
Schmitt, Anthony D., et al. A compendium of chromatin contact maps reveals spatially active regions in the human genome. Cell Reports 17, 2042–2059 (2016).
Li, M., et al. Integrative functional genomic analysis of human brain development and neuropsychiatric risks. Science 362, eaat7615 (2018).
Wang, K., Li, M. & Hakonarson, H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Research 38, e164-e164 (2010).
Rentzsch, P., Witten, D., Cooper, G.M., Shendure, J. & Kircher, M. CADD: predicting the deleteriousness of variants throughout the human genome. Nucleic Acids Research 47, D886-D894 (2019).
Boyle, A.P., et al. Annotation of functional variation in personal genomes using RegulomeDB. Genome Research 22, 1790–1797 (2012).
Kircher, M., et al. A general framework for estimating the relative pathogenicity of human genetic variants. Nature Genetics 46, 310–315 (2014).
Dong, S., et al. Annotating and prioritizing human non-coding variants with RegulomeDB v.2. Nature Genetics 55, 724–726 (2023).
de Leeuw, C.A., Mooij, J.M., Heskes, T. & Posthuma, D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Computational Biology 11, e1004219 (2015).
Liberzon, A., et al. The Molecular Signatures Database Hallmark Gene Set Collection. Cell Systems 1, 417–425 (2015).
Zhu, Z., et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nature Genetics 48, 481–487 (2016).
de Klein, N., et al. Brain expression quantitative trait locus and network analyses reveal downstream effects and putative drivers for brain-related diseases. Nature Genetics 55, 377–388 (2023).
Sun, B.B., et al. Plasma proteomic associations with genetics and health in the UK Biobank. Nature 622, 329–338 (2023).
Ferkingstad, E., et al. Large-scale integration of the plasma proteome with genetics and disease. Nature Genetics 53, 1712–1721 (2021).
Barbeira, A.N., et al. Exploring the phenotypic consequences of tissue specific gene expression variation inferred from GWAS summary statistics. Nature Communications 9, 1825 (2018).
Barbeira, A.N., et al. Integrating predicted transcriptome from multiple tissues improves association detection. PLOS Genetics 15, e1007889 (2019).
Gandal, M.J., et al. Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science 362, eaat8127 (2018).
Cuéllar-Partida, G., et al. Complex-Traits Genetics Virtual Lab: A community-driven web platform for post-GWAS analyses. bioRxiv, 518027 (2019).
Bulik-Sullivan, B., et al. An atlas of genetic correlations across human diseases and traits. Nature Genetics 47, 1236–1241 (2015).
Bulik-Sullivan, B.K., et al. LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nature Genetics 47, 291–295 (2015).
Auton, A., et al. A global reference for human genetic variation. Nature 526, 68–74 (2015).
Dennis, J.K., et al. Clinical laboratory test-wide association scan of polygenic scores identifies biomarkers of complex disease. Genome Medicine 13, 6 (2021).
Ge, T., Chen, C.-Y., Ni, Y., Feng, Y.-C.A. & Smoller, J.W. Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nature Communications 10, 1776 (2019).
Roden, D.M., et al. Development of a large-scale de-identified DNA biobank to enable personalized medicine. Clinical Pharmacology & Therapeutics 84, 362–369 (2008).
Purcell, S., et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. American Journal of Human Genetics 81, 559–575 (2007).
McCarthy, S., et al. A reference panel of 64,976 haplotypes for genotype imputation. Nature Genetics 48, 1279–1283 (2016).
Verma, A., et al. The Penn Medicine BioBank: Towards a genomics-enabled learning healthcare system to accelerate precision medicine in a diverse population. Journal of Personalized Medicine, Vol. 12 (2022).
Fuchsberger, C., Abecasis, G.R. & Hinds, D.A. minimac2: Faster genotype imputation. Bioinformatics 31, 782–784 (2015).
Das, S., et al. Next-generation genotype imputation service and methods. Nature Genetics 48, 1284–1287 (2016).
Kember, R.L., et al. Phenome-wide association analysis of substance use disorders in a deeply phenotyped sample. Biological Psychiatry 93, 536–545 (2023).
Pierucci-Lagha, A., et al. Diagnostic reliability of the Semi-structured Assessment for Drug Dependence and Alcoholism (SSADDA). Drug and Alcohol Dependence 80, 303–312 (2005).
Corsello, S.M., et al. The Drug Repurposing Hub: A next-generation drug library and information resource. Nature Medicine 23, 405–408 (2017).
Hatoum, A.S., et al. Multivariate genome-wide association meta-analysis of over 1 million subjects identifies loci underlying multiple substance use disorders. Nature Mental Health 1, 210–223 (2023).
Bell, N., Uffelmann, E., van Walree, E., de Leeuw, C. & Posthuma, D. Using genome-wide association results to identify drug repurposing candidates. medRxiv, 2022.2009.2006.22279660 (2022).
Cannon, M., et al. DGIdb 5.0: Rebuilding the drug–gene interaction database for precision medicine and drug discovery platforms. Nucleic Acids Research 52, D1227-D1235 (2024).
Hemani, G., et al. The MR-Base platform supports systematic causal inference across the human phenome. eLife 7, e34408 (2018).
Kurilshikov, A., et al. Large-scale association analyses identify host factors influencing human gut microbiome composition. Nature Genetics 53, 156–165 (2021).
Lopera-Maya, E.A., et al. Effect of host genetics on the gut microbiome in 7,738 participants of the Dutch Microbiome Project. Nature Genetics 54, 143–151 (2022).
Hemani, G., Tilling, K. & Davey Smith, G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLOS Genetics 13, e1007081 (2017).
Bostick, M., et al. UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science 317, 1760–1764 (2007).
Jourdon, A., Scuderi, S., Capauto, D., Abyzov, A. & Vaccarino, F.M. PsychENCODE and beyond: Transcriptomics and epigenomics of brain development and organoids. Neuropsychopharmacology 46, 70–85 (2021).
Markham, A. & Keam, S.J. Selumetinib: First approval. Drugs 80, 931–937 (2020).
Löwe, B., et al. Persistent physical symptoms: Definition, genesis, and management. The Lancet 403, 2649–2662 (2024).
Burton, C., et al. Functional somatic disorders: Discussion paper for a new common classification for research and clinical use. BMC Medicine 18, 34 (2020).
Li, C.N., Keay, K.A., Henderson, L.A. & Mychasiuk, R. Re-examining the mysterious role of the cerebellum in pain. The Journal of Neuroscience 44, e1538232024 (2024).
Moulton, E.A., Schmahmann, J.D., Becerra, L. & Borsook, D. The cerebellum and pain: Passive integrator or active participator? Brain Research Reviews 65, 14–27 (2010).
Szabo, E., Timmers, I., Borsook, D., Simons, L.E. & Sieberg, C.B. Altered anterior insula functional connectivity in adolescent and young women with endometriosis-associated pain: Pilot resting-state fMRI study. European Journal of Paediatric Neurology 41, 80–90 (2022).
Kim, J., et al. Somatotopically specific primary somatosensory connectivity to salience and default mode networks encodes clinical pain. Pain 160(2019).
van Ettinger-Veenstra, H., et al. Chronic widespread pain patients show disrupted cortical connectivity in default mode and salience networks, modulated by pain sensitivity. Journal of Pain Research 12, 1743–1755 (2019).
Pinto, A.M., et al. Emotion regulation and the salience network: A hypothetical integrative model of fibromyalgia. Nature Reviews Rheumatology 19, 44–60 (2023).
Icenhour, A., et al. Brain functional connectivity is associated with visceral sensitivity in women with Irritable Bowel Syndrome. NeuroImage: Clinical 15, 449–457 (2017).
Chen, Q., et al. Dysfunction of cortical GABAergic neurons leads to sensory hyper-reactivity in a Shank3 mouse model of ASD. Nature Neuroscience 23, 520–532 (2020).
Cheetham, S.W., Faulkner, G.J. & Dinger, M.E. Overcoming challenges and dogmas to understand the functions of pseudogenes. Nature Reviews Genetics 21, 191–201 (2020).
Ze, X., Duncan, S.H., Louis, P. & Flint, H.J. Ruminococcus bromii is a keystone species for the degradation of resistant starch in the human colon. The ISME Journal 6, 1535–1543 (2012).
Jandhyala, S.M., et al. Altered intestinal microbiota in patients with chronic pancreatitis: implications in diabetes and metabolic abnormalities. Scientific Reports 7, 43640 (2017).
Kowalska-Duplaga, K., et al. Differences in the intestinal microbiome of healthy children and patients with newly diagnosed Crohn’s disease. Scientific Reports 9, 18880 (2019).
David, L.A., et al. Diet rapidly and reproducibly alters the human gut microbiome. Nature 505, 559–563 (2014).
Meyer, W., Bühring, K.U., Steiner, K., Ungethüm, W. & Schnurr, E. Pharmacokinetics and first clinical experiences with an antihypertensive dopamine (DA2) agonist. European Heart Journal 13, 121–128 (1992).
Stankey, C.T., et al. A disease-associated gene desert directs macrophage inflammation through ETS2. Nature 630, 447–456 (2024).
Ji, R.-R., Gereau, R.W., Malcangio, M. & Strichartz, G.R. MAP kinase and pain. Brain Research Reviews 60, 135–148 (2009).
Khan, Y., et al. Combining transdiagnostic and disorder-level GWAS enhances precision of psychiatric genetic risk profiles in a multi-ancestry sample. medRxiv, 2024.2005.2009.24307111 (2024).

Yes there is potential Competing Interest. Dr. Kranzler is a member of advisory boards for Dicerna Pharmaceuticals, Sophrosyne Pharmaceuticals, Enthion Pharmaceuticals, Clearmind Medicine, and Altimmune; a consultant to Sobrera Pharmaceuticals; the recipient of research funding and medication supplies for an investigator-initiated study from Alkermes; and a member of the American Society of Clinical Psychopharmacology’s Alcohol Clinical Trials Initiative, which was supported in the last three years by Alkermes, Dicerna, Ethypharm, Lundbeck, Mitsubishi, Otsuka, and Pear Therapeutics; Drs. Gelernter and Kranzler hold U.S. patent 10,900,082 titled: "Genotype-guided dosing of opioid agonists," issued 26 January 2021.

Download PDF

Version 1

posted

You are reading this latest preprint version

Multivariate, Multi-omic Analysis in 799,429 Individuals Identifies 134 Loci Associated with Somatoform Traits

Status:

Version 1

Abstract

Figures

Introduction

Methods

Summary Statistics

Genomic Structural Equation Modeling

Gene Mapping and Functional Annotation

MAGMA Gene-Based Analyses

eQTL and pQTL Association Analyses

Transcriptome-Wide Association Studies

Polygenic Overlap with Psychopathology

Genetic Correlations with Publicly Available GWAS

Lab- and Phenome-Wide Association Studies

Drug Repurposing

Mendelian Randomization with Gut Microbiome Taxa

Results

Gene Mapping and Functional Annotation

Gene-Based Enrichment

SMR with Brain eQTLs and Blood Plasma pQTLs

Transcriptome-Wide Association Analyses

Polygenicity, Discoverability, and Polygenic Overlap with Psychopathology

Genetic Correlations

Lab- and Phenome-Wide Association Scans

Drug Repurposing

Potentially Causal Effects of the Gut Microbiome

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1