A genome-wide cross-trait analysis characterizes the shared genetic architecture between lung and gastrointestinal diseases

Abstract

Lung and gastrointestinal diseases often occur together, leading to more adverse health outcomes than when a disease of one of these systems occurs alone. However, the potential genetic mechanisms underlying lung-gastrointestinal comorbidities remain unclear. Here, we leverage lung and gastrointestinal trait data from individuals of European, East Asian and African ancestries, to perform a large-scale genetic cross trait analysis, followed by functional annotation and Mendelian randomization analysis to explore the genetic mechanisms involved in the development of lung-gastrointestinal comorbidities. Notably, we find significant genetic correlations between 27 trait pairs among the European population. The highest correlation is between chronic bronchitis and peptic ulcer disease. At the variant level, we identify 42 candidate pleiotropic genetic variants (3 of them previously uncharacterized) in 14 trait pairs by integrating cross-trait meta-analysis, fine-mapping and colocalization analyses. We also find 66 candidate pleiotropic genes, most of which were enriched in immune or inflammatory response-related activities. Causal inference approaches result in 4 potential lung-gastrointestinal associations. Introducing the gut microbiota as a variable establishes a relationship between the genus Parasutterella, gastro-oesophageal reflux disease and asthma. In summary, our findings highlight the genetic relationship between lung and gastrointestinal diseases, providing insights into the genetic mechanisms underlying the development of lung gastrointestinal comorbidities.

Fig. 1. Overview of the study design.

a Summary of the comprehensive pleiotropic analyses from different perspectives for multiple lung and gastrointestinal diseases in individuals of European, East Asian, and African ancestry. The genetic correlations at the global and local levels in each population were firstly examined. Created in BioRender. You, D. (2025) https://BioRender.com/j53h646. b Cross-trait meta-analysis and gene-based analyses were conducted to identify causal pleiotropic variants and genes. The functional annotation, TWAS, PWAS, and gene-environment interaction analyses were subsequently performed to explore biological functions. c Bi-directional MR and mediation analyses were performed to identify putative causal relationships and further clarify the shared biological mechanism between lung and gastrointestinal diseases. Created in BioRender. You, D. (2025) https://BioRender.com/j53h646. EUR: European. EAS: East Asian. AFR: African. AB: acute bronchitis. CB: chronic bronchitis. COPD: chronic obstructive pulmonary disease. FEV1: forced expiratory volume in 1 s. FVC: forced vital capacity. FEV1/FVC: FEV1/FVC ratio. PEF: peak expiratory flow. ILD: interstitial lung disease. IPF: idiopathic pulmonary fibrosis. LUAD: lung adenocarcinoma. LUSC: lung squamous cell carcinoma. SCLC: small cell lung carcinoma. LCES: lung cancer in ever smokers. LCNS: lung cancer in never smokers. Bact-pneumo: bacterial pneumoniae. Viral-pneumo: viral pneumonia. CP: colon polyp. CRC: colorectal cancer. DD: diverticular disease. GORD: gastro-oesophageal reflux disease. IBD: inflammatory bowel disease. CD: Crohn’s disease. UC: ulcerative colitis. IBS: irritable bowel syndrome. PUD: peptic ulcer disease. TWAS: transcriptome-wide association study. PWAS: proteome-wide association study. SNP: single-nucleotide polymorphism. MR: Mendelian randomisation. IV: instrumental variable.

Fig. 2. Summary of the genetic associations between 27 lung-gastrointestinal trait pairs.

a UpSet plot showing SNP-based heritability on the liability scale for each trait computed using LDSC (left bars, SNP-based liability heritability range: 0.02-0.21; standard error (SE) range: 0.002–0.02), global genetic correlations estimated using cross-trait LDSC (red bars, global genetic correlation range: 0.19–0.65; SE range: 0.04–0.07) and local levels computed using SUPERGNOVA [24 (17 unique) significant local genetic correlations] across 27 lung-gastrointestinal trait pairs. Lung-gastrointestinal trait pairs are indicated by blue dots. b Genome-wide map of 42 candidate pleiotropic genetic variants (dots, see also Supplementary Data 8) and the 66 corresponding candidate pleiotropic genes (see also Supplementary Data 15) for each gastrointestinal disease as a Circos plot. The region shared by multiple trait pairs is denoted by black text and a black-dotted bordered rectangle. Bonferroni correction was applied, and all reported P values were two-sided unless stated otherwise. LDSC: linkage disequilibrium score regression. CP: colon polyp. DD: diverticular disease. GORD: gastro-oesophageal reflux disease. IBS: irritable bowel syndrome. PUD: peptic ulcer disease. Bact-pneumo: bacterial pneumoniae. CB: chronic bronchitis. COPD: chronic obstructive pulmonary disease. IPF: idiopathic pulmonary fibrosis. LC: lung cancer. CRC: colorectal cancer. LUSC: lung squamous cell carcinoma. Source data were provided as a Source Data file.

Fig. 3. Functional annotation of pleiotropic genes across 14 lung-gastrointestinal trait pairs.

The heatmaps display functional annotations on the x-axis and 14 lung-gastrointestinal trait pairs on the y-axis, with significant cells filled in colour. The bars represented the number of trait pairs for each annotation item and the number of significant annotation items for each trait pair, respectively. Significantly enriched items were determined with Bonferroni-corrected P ≤ 0.05 for GSEA [normalised enrichment score > 2, GO terms and KEGG terms], TSEA, and CSEA, as well as FDR-corrected P ≤ 0.05 for TWAS and PWAS. All annotations were evaluated against a two-sided alternative hypothesis. GO and KEGG pathway enrichment analyses revealed that 24 unique biological processes related to immune or inflammatory response related activities were enriched (see also Supplementary Data 12). TSEA using the deTS method revealed 5 significantly enriched tissues (see also Supplementary Data 13). The top 5 enriched immune cell types from CSEA calculated by WebCSEA are shown (see also Supplementary Data 14). TWAS and PWAS were performed using FUSION based on 49 normal tissues from GTEx V8 and the plasma proteome built from the ARIC study, respectively (see also Supplementary Data 16-17). GO: Gene Ontology. KEGG: Kyoto Encyclopedia of Genes and Genomes. GSEA: gene set enrichment analysis. TSEA: tissue-specific enrichment analysis. CSEA: cell-type specific enrichment analysis. TWAS: transcriptome-wide association study. PWAS: proteome-wide association study. GTEx: Genotype-Tissue Expression project. CP: colon polyp. DD: diverticular disease. GORD: gastro-oesophageal reflux disease. IBS: irritable bowel syndrome. CB: chronic bronchitis. PUD: peptic ulcer disease. LUSC: lung squamous cell carcinoma. CRC: colorectal cancer. Source data were provided as a Source Data file.

Fig. 4. Causal inference of lung and gastrointestinal diseases.

a Significant bidirectional MR estimates (see also Supplementary Data 22). Data were presented as forest plots, with the OR shown as dots and the 95%CI as line segments, of the following samples per direction: asthma (n = 360,838) to GORD (n = 456,327); GORD (n = 385,276) to asthma (n = 1,376,071), CB (n = 450,422), or COPD (n = 995,917). Significance was defined as an FDR-corrected P ≤ 0.05 for IVW and a P > 0.05 for the Egger intercept, evaluated against a two-sided alternative hypothesis. Created in BioRender. You, D. (2025) https://BioRender.com/f18p045. b Mediation analysis based on IVW estimates (see also Supplementary Data 24-25). ‘Direct effect’ indicates the effect of the exposure on the outcome after adjusting for the mediator. The ‘mediation effect’ indicates the effect of the gut microbiota on the lung disease through the gastrointestinal disease, computed as the product of α (the IVW-estimated effect size of the exposure on the mediator) and β (effect size of the mediator on the outcome). The ‘mediation proportion’ was computed as the mediation effect divided by the total effect. The standard error (SE) and 95%CI of the mediation effect were calculated using the delta method. c Immune response in the pathogenesis of the Parasutterella-GORD-asthma association. Li et al. suggested that the “gut microbiota-inosine-PPARγ” axis plays an important role in the pathogenesis of chronic inflammatory disease of the intestine by regulating an inflammatory response (i.e., TNF-α and IL-6) through PPARγ signalling in colon epithelial cells, which further inhibited the immune response, leading to the increased the risk of GORD, as well as asthma. Created in BioRender. You, D. (2025) https://BioRender.com/f18p045. All reported P values were two-sided, unless stated otherwise. MR: Mendelian randomisation. IVW: inverse variance weighted method. MR-RAPS: MR robust adjusted profile score. MR-PRESSO: MR pleiotropy residual sum and outlier. OR: odds ratio. CI: confidence interval. GORD: gastro-oesophageal reflux disease. CB: chronic bronchitis. CP: colon polyp. COPD: chronic obstructive pulmonary disease. IL: interleukin. TNF-α: tumour necrosis factor-α. PPARγ: peroxisome proliferator activated receptor γ. Source data were provided as a Source Data file.