Discovering causal genes and comorbidities for asthma

doi:10.21203/rs.3.rs-3555707/v1

Download PDF

Research Article

Discovering causal genes and comorbidities for asthma

https://doi.org/10.21203/rs.3.rs-3555707/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Asthma is a heterogeneous phenotype that is often associated with other phenotypes. Identifying the genetic mediators that modulate the interaction between asthma and asthma-associated conditions will help inform our understanding of asthma heterogeneity. Here, we used Mendelian randomisation to identify asthma causal genes and their modifier spatial eQTLs within lung and whole blood-specific gene regulatory networks (GRNs), which integrate information on spatial genome organisation with tissue-specific expression quantitative trait loci (eQTL) data. Subsequently, we located the asthma-causal genes in the tissue-specific GRNs to define a putative asthma GRN and identified curated protein interaction partners occurring up to 4 edges (levels) away from the asthma GRN (level 0). We then queried the GWAS Catalog with the spatial eQTLs regulating level 0-4 genes to identify the GWAS traits enriched at each level (hypergeometric test; FDR≤0.05). This identified 113 traits significantly enriched in the regulatory space proximal to asthma, 106 of which had known associations with asthma (e.g., systemic lupus erythematosus and age-related macular degeneration) and seven traits whose association with asthma is yet to be confirmed. Importantly, our analysis identifies the genes and SNPs that modulate the interaction between asthma and asthma-associated traits by identifying the direct and indirect protein interacting partners of asthma causal genes. Finally, we highlight the druggable genes identified in our analysis, thereby providing new drug-repurposing opportunities for asthma.

asthma

Mendelian randomisation

gene regulatory networks

SNP function

comorbidity

expression quantitative trait loci

network analysis

Asthma is a chronic airway disease affecting over 340 million people worldwide [1]. It is characterised by hyperresponsiveness, obstruction, and remodelling of the airways, along with mucus overproduction [2]. Deciphering the genetic architecture of asthma is complicated by the disease’s heterogeneity; not only does asthma have different presentations, but it is also associated and comorbid with various other conditions, including respiratory, infectious, autoimmune, and metabolic diseases [3–5]. While genome-wide association studies (GWAS) have associated more than 800 single nucleotide polymorphisms (SNPs) with asthma [6], functional annotation of the SNPs and understanding the role they play in modulating the interaction of asthma with other diseases remains an ongoing challenge [7].

The exploration of neighbourhoods within protein-protein interaction networks (PPIN) of complex diseases has identified modules of highly connected genes having similar biological functions (8) and protein sets that are shared between predefined comorbid conditions (e.g., amyotrophic lateral sclerosis and frontotemporal dementia [8] and asthma and chronic obstructive pulmonary disease (COPD)) [9]. These network analyses are based on the premise that proteins within a PPIN are not randomly distributed; instead, they cluster in biologically comparable modules or pathways [9–12]. Furthermore, proteins linked to biologically similar diseases, comorbidities, or diseases that share symptoms tend to be in close proximity to each other within a PPIN [13]. Recently, we demonstrated that tissue-specific gene regulatory networks (GRNs), comprised of spatial expression quantitative trait loci (eQTLs) and the genes they regulate, can be used to identify putative gene targets for asthma-associated SNPs, the phenotypes that directly and indirectly interact with asthma, and the eQTLs and genes that bridge the interaction between asthma and interacting phenotypes [7]. Importantly, this identification can be completed without the need to pre-define the interacting conditions.

When conducting network analysis to identify biologically associated phenotypes, the selection of seed genes becomes a critical factor, and these genes should reflect the biology of the index condition of interest [14]. Mendelian randomisation (MR) is one method that can be used to identify genes that are causally related to and thus reflect the biology of index conditions – including asthma [15]. Here, we used MR to identify asthma-causal genes within a tissue-specific lung GRN (L-GRN) and whole-blood GRN (B-GRN). We characterise these genes and use them to prioritise potential driver cell types for asthma and to define a putative asthma GRN. Traits enriched in the regulatory space proximal to the asthma GRN are hypothesised to be comorbidities and biomarkers associated with asthma sub-types. Subsequently, we identify 113 such traits and further use patient hospital records to confirm which are observed asthma comorbidities within the New Zealand population. Our finding of unique and shared causal genes within blood and lung highlights the potential existence of asthma aetiologies dependent upon an individual's inherited germline variation – opening the way to more individualised therapeutic approaches in the future.

2.1. Building the lung and whole blood gene regulatory networks

Spatial lung (L-GRN) and whole blood GRN (B-GRN) were constructed as described in Zaied et al. (Fig. 1a) [7]. Briefly, all SNPs genotyped in GTEx (n = 40x10⁶, MAF ≥ 0.05; dbGaP accession: phs000424.v8.p2; project number: #22937) were run through the CoDeS3D algorithm [16]. CoDeS3D subsequently screens for tissue-specific spatial eQTLs by combining GTEx lung and blood eQTL data with Hi-C libraries of primary lung and blood cell lines. This results in the construction of the L-GRN and B-GRN, which consist of spatial eQTLs (eQTLs that physically interact with and regulate the expression of their target genes) identified within the tissue of interest, along with their target genes expressed in the same tissue.

2.2. Mendelian randomisation analysis

Mendelian randomisation (MR) was performed to identify genes causally linked with asthma in the L-GRN and B-GRN using the TwoSampleMR package (v0.5.6, https://mrcieu.github.io/TwoSampleMR/) [17]. For the MR analysis, spatial eQTL (p-value < 1x10^− 5) -gene pairs from the L-GRN were used as exposure data and the Valette et al.[15] asthma GWAS was used as outcome (i.e., disease) data. In MR, three assumptions must be met for genetic variants to be considered “instrumental variables” (IVs) to estimate the causal effects of an exposure on an outcome: 1) the genetic variant is associated with the exposure (relevance or exogeneity assumption), 2) the genetic variant is not associated with any confounding factors that might affect the exposure-outcome relationship (independence assumption), and 3) the variant doesn’t have any independent or indirect effect on the outcome (exclusion restriction assumption). Linkage disequilibrium (LD) clumping was performed to retain independent instruments using a window of 10,000 kb and an r² cutoff of < 0.001. The exposure and outcome data were harmonised so that the effect of a spatial eQTL on an outcome and exposure is relative to the same allele. MR was performed using the Wald ratio method for single-instrument exposures and the inverse variance weighted (IVW) method for exposures with multiple instruments. Target genes passing the Bonferroni correction significance threshold (p < 0.05/number of exposures) and sensitivity analysis (heterogeneity test [Q_pval < 0.05], only performed for exposures with multiple IVs) were considered statistically significant causal asthma genes. The same analysis was performed using the B-GRN (Fig. 1b). Custom tracks on the UCSC Genome Browser (https://genome.ucsc.edu/cgi-bin/hgCustom, GRCh38) were used to visualise spatial eQTLs and their target genes.

2.3. Biotype annotation of causal asthma genes

Biotype annotation (e.g., protein-coding, pseudogenes, lincRNA, etc.) of the causal asthma genes in lung and blood was achieved using the GENCODE database (http://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_26/gencode.v26.annotation.gtf.gz, accessed on 05/06/2021).

2.4. Gene Ontology analysis

Gene Ontology (GO) enrichment was performed on the causal asthma genes in lung and blood using the gost() module of the R package gprofiler2 [18]. Genes were tested for enrichment within KEGG pathways [19]. Multiple testing correction was performed, and pathways having a Benjamini-Hochberg[20] adjusted p-value ≤ 0.05 were deemed significant.

2.5. Identification of immune cell types and tissues having shared regulatory patterns of causal asthma genes

Spatial eQTL-gene pairs identified using MR were mapped onto immune cell types from the Database of Immune Cell Expression, Expression of quantitative trait loci and Epigenomics (DICE, https://dice-database.org/downloads/, accessed 09/02/2021) [21]. The pheatmap package (version 1.0.12) was used to perform unsupervised hierarchical clustering of eQTL effect sizes (beta) obtained from DICE across the 15 DICE immune cell types.

2.6. Identification of asthma-associated traits

Traits associated with asthma were identified using the multimorbid3D algorithm (https://github.com/Genome3d/multimorbid3D). Briefly, the L-GRN and genes identified as causal for asthma in the lung using MR were used as input to the multimorbid3D algorithm. Accordingly, the base asthma L-GRN, or level 0, consisted of genes identified as causal for asthma in the lung and their regulatory spatial eQTLs (including LD SNPs r² ≥ 0.8, window = ± 5000 bases, population = 1,000 genomes phase 3 European super-population). Curated proteins from STRING[22] (the top five highest scoring interacting proteins having a confidence score of ≥ 0.900 were considered; accessed 29/04/2022) and PROPER[23] that, directly and indirectly, interact with level 0 were then identified by expanding four edges or levels outwards from level 0 (Supplementary Data 1 and 2). Accordingly, proteins on levels 1–4 were identified where level 1 proteins share 1 edge with level 0 proteins (direct interaction partners of level 0 proteins) and level 2 proteins share 2 edges with level 0 proteins (indirect interaction partners of level 0 proteins), etc. The L-GRN was then queried to identify all spatial eQTLs that regulate the expression of level 0–4 genes. Traits associated with the identified spatial eQTLs were then retrieved by querying the GWAS Catalog. Of those, traits enriched at each level were identified using the hypergeometric enrichment test and significantly enriched traits were identified by multiple testing corrections (Bonferroni method, FDR < 0.05). Finally, a Monte Carlo simulation was carried out to identify asthma-specific traits. This was done by repeating the entire analysis 500 times using a random set of genes equal to the number of genes identified in level 0. Significant spatial eQTLs, proteins, and traits having an adjusted p-value < 0.05 were identified (Fig. 1c).

The same analysis was repeated using the B-GRN to identify asthma-associated traits in the blood with the asthma B-GRN (causal asthma genes in blood) serving as seed genes (Supplementary Data 3 and 4).

2.7. DisGeNet and comorbidity analysis

DisGeNet [24] was used to identify the gene-disease associations (GDAs) of L-GRN and B- GRN genes identified in levels 0–4. Only GDAs derived from CURATED databases were used in this analysis. The identified diseases, reported in UMLS CUI identifiers, were converted to ICD-10 terms using the disease mapping table (accessed 2/07/2022). Disease mapping tables from the Hospital Pricing Authority (IHPA) were used to convert ICD-10 disease terms to ICD-10-AM (https://www.ihpa.gov.au/publications/icd-10-am-and-achi-mapping-tables, icd-10_2016_to_icd-10-am_eleventh_edition_0.zip, accessed 7/8/2022; Fig. 1d). The ComoRbidity R package was used to identify comorbidities of asthma using hospital records of 1,119,021 females and 2,051,658 males admitted between 31/12/2015 and 01/01/2021 in New Zealand. Hospital records were obtained from the Integrated Data Infrastructure (Fig. 1e). Detailed method is described in Zaied et al. [7].

2.8. Ethics

The Auckland Health Research Ethics Committee (approval AH22495) and Statistics New Zealand (project number MAA2020-63) reviewed and approved the use of the Integrated Data Infrastructure within Stats NZ Data Lab.

3.1. eQTL analysis identifies spatially constrained eQTLs in the lung and blood

Disease-associated SNPs are typically located within non-coding DNA, can participate in long-range chromatin interactions, and have small effect sizes [25]. These factors make attributing function to disease-associated SNPs an ongoing challenge [26]. We combined chromatin interaction data (Hi-C) and eQTL data from GTEx to build spatially constrained GRNs in which putative functions are assigned to SNPs by associating them with their target genes (Fig. 1a) (17) The resulting L-GRN has 740,028 spatial eQTLs regulating the expression of 15,855 genes expressed in the lung, constituting 873,133 spatial eQTL-gene interactions. The B-GRN has 1,077,379 spatial eQTLs regulating the expression of 14,871 genes expressed in whole blood, making a total of 1,713,885 spatial eQTL-gene interactions. The L-GRN and B-GRN are described extensively in Zaied et al. [7]

3.2. Mendelian randomisation identifies genes causally associated with asthma in the lung and blood

MR infers the causal effect of exposure (genes) on the disease outcomes (development of asthma) using genetic variants as IVs [17]. We conducted a two-sample MR analysis to test whether L-GRN and B-GRN genes were causally associated with the development of asthma using spatial eQTLs as IVs (Fig. 1b). Among exposures with single IVs, the Wald ratio method identified 65 genes regulated by 64 spatial eQTLs as causally associated with asthma in the L-GRN (Fig. 2a; Supplementary Data 5). No exposures with multiple IVs were identified to be causally linked to asthma in the L-GRN. Of the 65 asthma causal genes in the lung, 43 were protein-coding, and 22 were “other” biotypes, e.g., antisense [n = 10] and pseudogenes [n = 6], etc., Fig. S1). These genes were enriched for 17 pathways in KEGG, including asthma, inflammatory bowel disease, and Th1 and Th2 cell differentiation (Fig. S2a). In the B-GRN, 63 genes regulated by 62 spatial eQTLs were identified using the Wald ratio method (Fig. 2b), and five genes regulated by 11 spatial eQTLs were identified using the IVW method (Fig. 2c; Supplementary Data 6). 53 of the identified asthma causal genes in blood were protein-coding, and 15 were non-protein-coding, e.g., antisense [n = 8] and processed transcripts [n = 3], etc., Fig. S1). The identified causal genes in blood were enriched for 12 KEGG pathways, including Th1 and Th2 cell differentiation, asthma, and Th17 cell differentiation (Fig. S2b).

Amongst the causal genes we identified, 69 were lung or blood-specific, and 32 were shared between the L-GRN and B-GRN. Of the shared genes, 27 had the same direction of effect on asthma risk in both tissues, and only three were regulated by the same spatial eQTLs across both tissues (Fig. 2d). Valette et al. identified 50 causal asthma genes in blood using non-spatial eQTLs as IVs [15]. Notably, our study independently replicated the causality of 19 of the causal genes: six in both the lung and blood (i.e., GSDMB, RERE, PGAP3, UNC13D, MEI1, and SCRN2), three exclusively in the lung (i.e., TMEM258, SLC22A5, and FADS1), and ten specific to the blood (i.e., IKZF3, SLC22A4, ING5, AHI1, SIK2, TEF, HOXB4, GNGT2, TLR10, and EFEMP2).

More than 200 genomic loci have been associated with the development of asthma through GWAS studies [6, 27]. The methods employed to identify the potential target genes of these loci include the nearest gene approach [28] and eQTL analysis [29, 30]. In this study, spatially constrained eQTLs were a critical component of identifying putative asthma causal genes. Of the identified causal genes in lung and blood, all were regulated in cis (spatial eQTLs located within +/-1 Mb of their target genes on the same chromosome) except for three eQTL-gene pairs (i.e., rs10167431-ATF2, rs9267123-HLA-DQB1, and rs61044849-RABL3) which were regulated by trans-intrachromosomal eQTLs (spatial eQTLs located > 1 Mb away on the same chromosome; Supplementary Data 6 and 7).

3.3. Seven protein-coding genes within the 17q12-21 locus are causal for asthma

49 of the 101 genes causally associated with asthma (48.5%) identified in lung and blood are located on chromosomes 17 (17 genes), five (17 genes), and six (15 genes) (Supplementary Data 5 and 6). Within chromosome 17, we identified seven protein-coding genes that are modifiable and causal for asthma in the 17q12-21 asthma locus (i.e., PGAP3, IKZF3, GSDMB, ORMDL3, GSDMA, MED24, and CASC3; Fig. 3). PGAP3, GSDMB, ORMDL3, and MED24 were identified as causal in both blood and lung, but are regulated by tissue-specific spatial eQTLs. Notably, some eQTLs coordinate the regulation of multiple genes (e.g., rs8067378 regulates both GSDMB and ORMDL3 in the blood). Furthermore, the eQTL site that regulates CASC3 in blood (rs3884477) is located within MED24, again providing the potential for coordination between these genes.

3.4 Shared regulatory patterns for causal asthma genes across immune cell types suggests overlapping impacts in the etiology of asthma

Various immune cell types have been associated with the development of asthma. Therefore, we mapped the causal asthma eQTL-gene pairs from lung and blood onto immune cell type-specific cis-eQTL data to prioritise potential driver immune cell types. Cell type-specific cis-eQTL data (not spatially constrained) was obtained from DICE across 15 immune cell types (i.e., innate immune cell types [natural killer cells [NK], classical monocytes, and non-classical monocytes [M2 cells]], naïve adaptive immune cells [naïve B, naïve CD4 + T cells, activated naïve CD4 + T cells [CD4_stim], naïve CD8 + T cells, activated naïve CD8 + T cells [CD8_stim], and naïve TREG cells], and CD4 + T memory subtypes [Memory TREG, TH1, TH 1/17 [TH star], TH17, TH2, and follicular helper T cells [TFH]]). Of the 63 causal asthma eQTL-gene pairs identified in the lung, 23 were identified as cis eQTL-gene pairs across 15 immune cell types (Fig. 4a). To distinguish the direction of effect in different cell populations, we performed unsupervised hierarchical clustering of effect sizes (beta) of cis-eQTLs across the cell types. In the lung, hierarchical clustering identified three main clusters of causal asthma eQTL-gene pairs: causal eQTL-gene pairs with A) negative, B) both positive and negative, and C) positive effect sizes (Fig. 4a). Additionally, four cell-type clusters (clusters 1–4) were enriched for various causal eQTL-gene pairs. For example, cluster 1 is the monocyte cluster and is enriched for causal eQTL-gene pairs from cluster A. Cluster 2 is primarily CD8 + and CD4 + T cells, and cell clusters 3 and 4 are predominantly CD4 + T memory subtypes.

Of the 65 single-IV causal eQTL-gene pairs identified in blood, 25 were also identified across the 15 DICE immune cell types (Fig. 4b). Hierarchical clustering identified three main clusters of causal asthma eQTL-gene pairs (A-C) and four cell-type clusters (clusters 1–4).

3.5. Asthma endotypes share regulatory mechanisms with the asthma GRNs

To identify traits that share regulatory connections with the causal asthma genes (i.e., level 0 traits), we retrieved the significantly enriched GWAS traits associated with spatial eQTLs (plus LD SNPs) targeting the asthma L-GRN (level 0). We identify 63 enriched traits (hypergeometric test, FDR ≤ 0.05 and 500 sets of Monte Carlo simulation) associated with spatial eQTLs (plus LD SNPs; n = 421) targeting 52 level 0 genes in lung. Convex biclustering of level 0 genes and traits revealed four distinct clusters (Fig. 5a). Cluster 1 includes genes known to play a role in cholesterol metabolism (e.g., HLA-DQA1, FADS1, and SLC22A5) and the spatial eQTLs regulating them were enriched for two cholesterol level traits. Cluster 2 is the metabolites cluster marked by two genes in 11q12.2 involved in fat metabolism (i.e., FADS1 and TMEM258). Cluster 3 includes immune response and allergy-related traits and includes D2HGDH, RERE, RP5-1115A15.1 and WDR36. Cluster 4 includes various endotypes of asthma, asthma medication, and other immune-mediated traits; it largely involves genes located within the 17q12-21 asthma locus (i.e., GSDMA, GSDMB, ORMDL3 and PGAP3; Fig. 5a).

Spatially constrained eQTLs and their LD SNPs within the asthma B-GRN were also analysed to identify traits sharing regulatory interactions with asthma in the blood (level 0 traits). 54 significantly enriched traits were associated with 512 SNPs regulating 57 level 0 genes in the blood. Convex biclustering of the level 0 genes and traits revealed five distinct clusters. Cluster 1 has two immunoglobulin-related traits and five HLA genes (i.e., HLA-DQA1, HLA-DQB1, HLA-DQB1-AS1, HLA-DQB2 and HLA-DRB1). The traits within cluster 2 include blood cell counts and genes related to cellular processing (i.e., MEI1, TEF, NDFIP1, and PMM1). Clusters 3 and 4 included immune-response and allergy-related traits and contained combinations of genes that resemble those in cluster 1. Cluster 5 included asthma endotypes and asthma medication-related traits. In addition to the HLA genes observed in cluster 1, cluster 5 also included genes within the 17q12-21 locus (i.e., GSDMB, ORMDL3, and IKZF3; Fig. 5b; Supplementary Data 7).

This brings the total number of traits enriched in level 0 to 88 across both tissues. Of these 88 traits, hepatitis C induced liver cirrhosis, IgA nephropathy, primary biliary cholangitis, Epstein Barr virus nuclear antigen 1 IgG levels and primary central nervous system lymphoma have not been previously identified as associating with asthma. The remaining traits we identified have known associations with asthma (e.g. systemic lupus erythematosus (SLE) [31], alphalinolenic acid [32], high density lipoprotein cholesterol [33], Immunoglobulin A vasculitis [34], eosinophil count [35], trans fatty acid levels [36], and Crohn’s disease [37]). Notably, 29 of the 88 level 0 traits are shared between both tissues; this can be attributed to the 24 shared asthma causal genes enriched at level 0 (hypergeometric test, FDR ≤ 0.05, and 500 sets of Monte Carlo simulations) as well as the multigenic nature of complex traits (Fig. 6a).

3.6. Additional traits are associated with asthma through indirect interactions with the asthma GRNs

Proteins do not work in isolation, and proteins associated with comorbid or biologically related diseases are more likely to interact within a PPIN [10, 38]. Therefore, we expanded four edges outwards from the asthma L-GRN (level 0; Fig. 6a) to identify its curated interacting protein partners (20,21)(Fig. 1c). This identified a total of 1,801 curated proteins across levels 1–4. Subsequently, Monte Carlo simulations reduced it to a set of 137 significant protein-coding genes. Level 1 included four interacting proteins and seven SNPs. At level 2, 123 genes associated with 231 SNPs were identified. While level 3 included 11 genes associated with 28 SNPs. No significant enrichment was identified at level 4 due to Monte Carlo simulation. To determine the traits enriched in the regulatory space proximal to the asthma L-GRN, we identified the enriched traits associated with SNPs that regulate level 1–3 proteins. In total, 34 enriched traits were identified across levels 1–3. Of those, seven traits, including asthma, were specific to levels 1–3 and unique to the lung (Fig. 6b). All of these had known associations with asthma, including keratinocyte cancer [39, 40], sphingolipid levels[41, 42] and amyotrophic lateral sclerosis (ALS) [43].

Performing this expansion analysis starting from the asthma B-GRN (Fig. 6a) identified 2,676 proteins over levels 1–4. Monte Carlo simulation reduced the total number of proteins to 43 protein-coding genes. Of these, 11 genes and 50 SNPs, ten genes and 25 SNPs, 16 genes and 30 SNPs, and six genes and 15 SNPs were identified at levels 1 through 4, respectively. SNPs that regulated level 1–4 genes were associated with 18 traits. Of those, 14 traits were blood-specific (Fig. 6c). Apart from aspartate aminotransferase platelet ratio index in high alcohol intake, all identified traits had reported associations with asthma. These include follicular lymphoma [44], shingles [45], Intracerebral hemorrhage [46], caffeine metabolism [47] and age-related macular degeneration [48].

A total of five enriched traits were shared between the lung and blood across levels 1–3 (Fig, 6d). Apart from nasopharyngeal carcinoma, all shared traits have previously been reported to be associated with asthma. These include chronic lymphocytic leukemia [49], IgE levels [50] and Hepatitis B [51, 52]. Importantly, this analysis also identifies the molecular mechanisms (i.e., genes and SNPs) that mediate the asthma-trait interaction. For instance (54,55)the enriched IgE levels trait is associated with three SNPs that regulate the expression of FCER1A and ACKR1, these have direct protein interactions with the asthma causal genes KCER1G and TNFRSF14, respectively (Fig. 6e).

3.7. Analysis of gene-disease associations corroborates identified asthma comorbidities observed in New Zealand patients.

Various disorders have been observed in asthmatics (e.g., diabetes and cardiovascular disease), consistent with the possible existence of common molecular mechanisms influencing their onset [53]. We identified 113 GWAS traits interacting with asthma (Fig. 6, Supplementary Data 7). However, some of these traits are not diseases (e.g., IgE levels). Therefore, we used DisGeNet [24] to identify the diseases associated with the genes in levels 0–4 across both tissues. This revealed 38 genes associated with 337 UMLS CUI identifiers in DisGeNet (Supplementary data 8), mapping to 71 three-digit ICD-10 disease terms and 72 curated UMLS CUI disease identifiers.

Hospitalisation records (n = 2,051,661; between 31/12/2015 and 01/01/2021) in New Zealand were analysed to identify comorbid disorders in patients with asthma (n = 26,781). The comorbidity analysis revealed 194 three-digit ICD-10-AM terms to have a statistical association with asthma (i.e., diseases that are more or less likely to be observed in the asthmatic population; q-value < 0.05; Supplementary data 9). Of those, 13 disease terms (mapping to 18 ICD-10-AM terms) were also identified in the GDA analysis. These diseases all had higher likelihoods of being comorbid with asthma (Fig. 7a). Moreover, the 13 disease terms were associated with nine genes identified in levels 0–4 across both tissues (Fig. 7b).

3.8. Genes bridging asthma-trait interactions are part of the druggable genome

To identify opportunities for drug repurposing, we investigated which of the genes identified in levels 0–4 (hypergeometric test, FDR ≤ 0.05 and 500 sets of Monte Carlo simulation) across both tissues are present in the druggable genome [54] and the drug-gene interaction database (DGIdb [55]). Of the 257 genes in levels 0–4, 61 are a part of the druggable genome, while 22 have known interactions with drugs in DGIdb (Supplementary data 10). Importantly, some of the identified genes are targets of known asthma drugs, including TSLP (Tezepelumab), IL2RA (daclizumab) and IL4R (Dupilumab).

In this study, we used MR to identify causal genes for asthma in the lung (n = 65) and blood (n = 68) GRNs. The expression of the causal genes we identified is modifiable across the lung, blood, and 15 DICE immune cell types through the action of non-spatial eQTLs (DICE) and spatial eQTLs (L-GRN and B-GRN). To our knowledge, this is the first study that identifies causal genes for asthma that are modifiable through spatial eQTLs. The identified causal asthma genes formed the seed genes in a PPIN analysis, which enabled the identification of 113 traits in the regulatory space proximal to asthma. Most of these traits have known associations with, or are comorbidities of, asthma. Importantly, our analysis also identified the genes and SNPs that mediate the observed asthma-trait interactions.

There are reported associations between asthma and AMD (OR = 2.34, 95% CI 1.22–4.46 [48]), a leading cause of blindness [48]. Chronic inflammation, systemic hypoxia, and complement activation are thought to be potential contributors to the observed asthma-AMD association [56]. However, the mechanism linking them remains unknown. Our analysis implicates eight SNPs and four genes (i.e., HLA-B, SKIV2L, TGFBR1, and TJP1) in the asthma-AMD interaction. Studies have found that specific subtypes of AMD exhibit altered expression of wound response genes, including members of the TGFβ pathway, in retinal pigment epithelial (RPE) cells, which have central roles in AMD development [57]. Consistent with this, Radeke et al. demonstrated that the targeted inhibition of TGFβR1 increases the lifespan of RPE cells [57]. Notably, defective TGFβR1 signaling is one of the biological perturbations increasing asthma risk [58]. Moreover, TGFβR1 (level 3) indirectly interacts with the asthma causal gene NDFIP1 (level 0). This interaction is mediated by NEDD4L (level 1) and SMAD7 (level 2), all of which have roles in TGFβ signaling [59, 60]. Thus, we contend that this is a potential mechanism through which association between asthma and AMD occur.

IgE levels are commonly found to be significantly elevated in patients with allergic asthma and are thus considered a potential biomarker of allergic asthma [61, 62]. Our analysis identifies the IgE trait enriched in level 1 across both tissues. Here, the level 0 asthma causal gene, FCER1G, interacts with FCER1A, which encodes the alpha subunit of the high-affinity IgE receptor (FCER1). When IgE binds to the FCER1 receptor, which is found on the surface of mast cells, basophils and antigen presenting cells [63, 64], it triggers the release of inflammatory mediators known to play a role in exacerbating asthma symptoms [64]. Accordingly, the interaction between FCER1G and FCER1A, which is the gene target of three IgE-associated SNPs, is one possible mechanism that links asthma with IgE levels. Similarly, the asthma causal gene TNFRSF14 identified in blood interacts with ACKR1, a crucial regulator of chemokine signalling expressed on the surface of erythrocytes and a subset of epithelial cells [65]. Our analysis identifies ACKRI as the gene target of the blood-specific SNP, rs13962, which is associated with IgE levels by GWAS (Fig. 6e). Notably, individuals with the FyES phenotype (reduced erythrocyte ACKR1) have increased serum IgE levels and higher risk for asthma [66, 67]. These results additionally highlight the importance of identifying tissue-specific genetic variation which, if taken into consideration can allow for the development of more precise personalised risk assessment.

Many of the traits identified in our analysis are not diseases (e.g., IgE levels). Therefore, it is difficult to connect them to phenotypes observed in patients. Accordingly, we curated GDAs using DisGeNet for the genes identified in levels 0–4. Of the identified GDAs, 13 were comorbidities observed in the asthmatic patients hospitalised in New Zealand (i.e., these diseases were prevalent in the asthmatic hospitalised population compared to the non-asthmatic patients). Notably, our analysis identifies nine genes that potentially bridge the observed comorbidities to asthma. This includes the circadian clock gene, CLOCK, which is associated with alcohol use disorders [68] and is thought to play a role in the marked day-night variation of asthma pathology [69].

This study is not without limitations. Firstly, our analysis uses data from healthy primary cell lines and does not investigate samples derived from asthmatic individuals, accordingly Hi-C interactions that are specific to diseased phenotypes may be missed by our analysis. Second, while modulating gene expression is one of the potential mechanisms for the realization of the genetic risks associated with asthma and its comorbidities, other mechanisms exist. For example, environmental and age-related mechanisms, variants altering other non-protein coding gene sequences (e.g., miRNAs), etc... can all theoretically alter the realization of inherited genetic risk. Third, measuring transcript levels alone is insufficient to explain changes in protein expression, as other mechanisms (e.g., post transcriptional processes) can alter expression levels. Fourth, the data available from DICE, while valuable, lacks information on some asthma relevant cell types (e.g. airway smooth muscle cells and eosinophils). However, despite these limitations, the finding that there are unique and shared causal genes within blood and lung should not be overlooked as it demonstrates the potential for the existence of different asthma aetiologies that are dependent upon an individual's inherited germline variation.

In conclusion, we identify genes that are causal for asthma in lung and blood and demonstrate their role in mediating the interaction between asthma and traits that interact with asthma. By integrating different levels of biological information, our de novo approach identifies traits associated with asthma development without a priori definition of the interacting phenotypes. Importantly, our findings highlight the molecular mechanisms (genes and variants) that modulate the identified asthma-trait interactions, thereby identifying opportunities for future drug development.

Data availability

Tissue-specific spatial GRNs are available from 10.6084/m9.figshare.20205644.v1 and 10.17608/k6.auckland.17067953.v1. Code is available at https://github.com/Genome3d/multimorbid3D.

Funding

This work was funded by the Dines Family Charitable Trust (J.M.O., R.E.Z., and S.G.).

Competing interests

The authors declare no competing interests.

Authors' contributions

R.E.Z. conceptualised, performed analysis, interpreted data, and wrote the manuscript. S.G. contributed to data analysis and manuscript revision. C.W. contributed to data interpretation and manuscript revision. J.M.O. directed the study, contributed to data interpretation and data analysis, and co-wrote the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We thank members of the Genomics and Systems Biology Group (Liggins Institute) for thoughtful discussions and suggestions. This work contains data from the Genotype-Tissue Expression (GTEx) Project, which was supported by the Common Fund of the Office of the Director of the National Institutes of Health and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS.

Materials & Correspondence

Correspondence and requests for materials should be addressed to J.M.O.

Vos T, Abajobir AA, Abbafati C, Abbas KM, Abate KH, Abd-Allah F, et al. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990-2016: A systematic analysis for the Global Burden of Disease Study 2016. The Lancet 2017;390. https://doi.org/10.1016/S0140-6736(17)32154-2.
King-Biggs MB. In the Clinic® asthma. Ann Intern Med 2019;171. https://doi.org/10.7326/AITC201910010.
Liljendahl MS, Sevelsted A, Chawes BL, Stokholm J, Bønnelykke K, Andersen ZJ, et al. Childhood asthma is associated with development of type 1 diabetes and inflammatory bowel diseases: a Danish nationwide registry study. Sci Rep 2022;12. https://doi.org/10.1038/s41598-022-26067-4.
Akmatov MK, Ermakova T, Holstiege J, Steffen A, von Stillfried D, Bätzing J. Comorbidity profile of patients with concurrent diagnoses of asthma and COPD in Germany. Sci Rep 2020;10. https://doi.org/10.1038/s41598-020-74966-1.
Liu L, Liu Y, Zhang X, Yuan YL, Chen ZH, Chen-Yu Hsu A, et al. Dyslipidemia Is Associated With Worse Asthma Clinical Outcomes: A Prospective Cohort Study. Journal of Allergy and Clinical Immunology: In Practice 2023;11. https://doi.org/10.1016/j.jaip.2022.11.037.
Demenais F, Margaritte-Jeannin P, Barnes KC, Cookson WOC, Altmüller J, Ang W, et al. Multiancestry association study identifies new asthma risk loci that colocalize with immune-cell enhancer marks. Nat Genet 2018;50:42–53.
Zaied RE, Fadason T, O’Sullivan JM. De novo identification of complex traits associated with asthma. Front Immunol 2023;14.
Mao Y, Kuo SW, Chen L, Heckman CJ, Jiang MC. The essential and downstream common proteins of amyotrophic lateral sclerosis: A protein-protein interaction network analysis. PLoS One 2017;12. https://doi.org/10.1371/journal.pone.0172246.
Maiorino E, Baek SH, Guo F, Zhou X, Kothari PH, Silverman EK, et al. Discovering the genes mediating the interactions between chronic respiratory diseases in the human interactome. Nat Commun 2020;11:1–14.
Oti M, Snel B, Huynen MA, Brunner HG. Predicting disease genes using protein-protein interactions. J Med Genet 2006;43. https://doi.org/10.1136/jmg.2006.041376.
Gratten J, Visscher PM. Genetic pleiotropy in complex traits and diseases: Implications for genomic medicine. Genome Med 2016;8. https://doi.org/10.1186/s13073-016-0332-x.
Sharma A, Menche J, Huang CC, Ort T, Zhou X, Kitsak M, et al. A disease module in the interactome explains disease heterogeneity, drug response and captures novel pathways and genes in asthma. Hum Mol Genet 2015;24:3005–20.
Tu Z, Argmann C, Wong KK, Mitnaul LJ, Edwards S, Sach IC, et al. Integrating siRNA and protein-protein interaction data to identify an expanded insulin signaling network. Genome Res 2009;19. https://doi.org/10.1101/gr.087890.108.
Doncheva NT, Kacprowski T, Albrecht M. Recent approaches to the prioritization of candidate disease genes. Wiley Interdiscip Rev Syst Biol Med 2012;4. https://doi.org/10.1002/wsbm.1177.
Valette K, Li Z, Bon-Baret V, Chignon A, Bérubé JC, Eslami A, et al. Prioritization of candidate causal genes for asthma in susceptibility loci derived from UK Biobank. Commun Biol 2021;4. https://doi.org/10.1038/s42003-021-02227-6.
Fadason T, Schierding W, Lumley T, O’Sullivan JM. Chromatin interactions and expression quantitative trait loci reveal genetic drivers of multimorbidities. Nat Commun 2018;9. https://doi.org/10.1038/s41467-018-07692-y.
Hemani G, Tilling K, Davey Smith G. Erratum: Correction: Orienting the causal relationship between imprecisely measured traits using GWAS summary data (PLoS genetics (2017) 13 11 (e1007081)). PLoS Genet 2017;13. https://doi.org/10.1371/journal.pgen.1007149.
Raudvere U, Kolberg L, Kuzmin I, Arak T, Adler P, Peterson H, et al. G:Profiler: A web server for functional enrichment analysis and conversions of gene lists (2019 update). Nucleic Acids Res 2019;47. https://doi.org/10.1093/nar/gkz369.
Kanehisa M, Furumichi M, Sato Y, Kawashima M, Ishiguro-Watanabe M. KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res 2023;51. https://doi.org/10.1093/nar/gkac963.
Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological) 1995;57. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x.
Schmiedel BJ, Singh D, Madrigal A, Valdovino-Gonzalez AG, White BM, Zapardiel-Gonzalo J, et al. Impact of Genetic Polymorphisms on Human Immune Cell Gene Expression. Cell 2018;175. https://doi.org/10.1016/j.cell.2018.10.022.
Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, et al. STRING v11: Protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 2019;47. https://doi.org/10.1093/nar/gky1131.
Johnson KL, Qi Z, Yan Z, Wen X, Nguyen TC, Zaleta-Rivera K, et al. Revealing protein-protein interactions at the transcriptome scale by sequencing. Mol Cell 2021;81:4091-4103.e9. https://doi.org/10.1016/j.molcel.2021.07.006.
Piñero J, Ramírez-Anguita JM, Saüch-Pitarch J, Ronzano F, Centeno E, Sanz F, et al. The DisGeNET knowledge platform for disease genomics: 2019 update. Nucleic Acids Res 2020;48. https://doi.org/10.1093/nar/gkz1021.
Visscher PM, Wray NR, Zhang Q, Sklar P, McCarthy MI, Brown MA, et al. 10 Years of GWAS Discovery: Biology, Function, and Translation. Am J Hum Genet 2017;101. https://doi.org/10.1016/j.ajhg.2017.06.005.
Fadason T, Farrow S, Gokuladhas S, Golovina E, Nyaga D, O’Sullivan JM, et al. Assigning function to SNPs: Considerations when interpreting genetic variation. Semin Cell Dev Biol, Elsevier; 2021.
Han Y, Jia Q, Jahani PS, Hurrell BP, Pan C, Huang P, et al. Genome-wide analysis highlights contribution of immune system pathways to the genetic architecture of asthma. Nat Commun 2020;11. https://doi.org/10.1038/s41467-020-15649-3.
Brodie A, Azaria JR, Ofran Y. How far from the SNP may the causative genes be? Nucleic Acids Res 2016;44. https://doi.org/10.1093/nar/gkw500.
Dekker J, Misteli T. Long-range chromatin interactions. Cold Spring Harb Perspect Biol 2015;7:a019356.
Aguet F, Barbeira AN, Bonazzola R, Brown A, Castel SE, Jo B, et al. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science (1979) 2020;369:1318–30. https://doi.org/10.1126/SCIENCE.AAZ1776.
Charoenngam N, Ponvilawan B, Wongtrakul W, Ungprasert P. Patients with asthma have a higher risk of systemic lupus erythematosus: a systematic review and meta-analysis. Clin Rheumatol 2021;40. https://doi.org/10.1007/s10067-020-05279-x.
Boskabady MH, Kaveh M, Shakeri F, Mohammadian Roshan N, Rezaee R. Alpha-linolenic acid ameliorates bronchial asthma features in ovalbumin-sensitized rats. Journal of Pharmacy and Pharmacology 2019;71. https://doi.org/10.1111/jphp.13094.
Schäfer T, Ruhdorfer S, Weigl L, Wessner D, Heinrich J, Döring A, et al. Intake of unsaturated fatty acids and HDL cholesterol levels are associated with manifestations of atopy in adults. Clinical and Experimental Allergy 2003;33. https://doi.org/10.1046/j.1365-2222.2003.01780.x.
Sürmeli Döven S, Ayçin G, Aydoğdu A. Association between Henoch-Schönlein Purpura with Gastrointestinal Involvement and Asthma in Children. Kocaeli Medical Journal 2021;10. https://doi.org/10.5505/ktd.2021.82713.
Tran TN, Khatry DB, Ke X, Ward CK, Gossage D. High blood eosinophil count is associated with more frequent asthma attacks in asthma patients. Annals of Allergy, Asthma and Immunology 2014;113. https://doi.org/10.1016/j.anai.2014.04.011.
Weiland SK, Von Mutius E, Hüsing A, Innes Asher M. Intake of trans fatty acids and prevalence of childhood asthma and allergies in Europe. Lancet 1999;353. https://doi.org/10.1016/S0140-6736(99)01609-8.
Hemminki K, Li X, Sundquist J, Sundquist K. Subsequent Autoimmune or Related Disease in Asthma Patients: Clustering of Diseases or Medical Care? Ann Epidemiol 2010;20. https://doi.org/10.1016/j.annepidem.2009.11.007.
Ideker T, Sharan R. Protein networks in disease. Genome Res 2008;18. https://doi.org/10.1101/gr.071852.107.
Cheng J, Zens MS, Duell E, Perry AE, Chapman MS, Karagas MR. History of allergy and atopic dermatitis in relation to squamous cell and basal cell carcinoma of the skin. Cancer Epidemiology Biomarkers and Prevention 2015;24. https://doi.org/10.1158/1055-9965.EPI-14-1243.
Yanik EL, Pfeiffer RM, Freedman DM, Weinstock MA, Cahoon EK, Arron ST, et al. Spectrum of immune-related conditions associated with risk of keratinocyte cancers among elderly adults in the United States. Cancer Epidemiology Biomarkers and Prevention 2017;26. https://doi.org/10.1158/1055-9965.EPI-17-0003.
Guo C, Sun L, Zhang L, Dong F, Zhang X, Yao L, et al. Serum sphingolipid profile in asthma. J Leukoc Biol 2021;110. https://doi.org/10.1002/JLB.3MA1120-719R.
Ono JG, Worgall TS, Worgall S. Airway reactivity and sphingolipids—implications for childhood asthma. Mol Cell Pediatr 2015;2. https://doi.org/10.1186/s40348-015-0025-3.
Turner MR, Goldacre R, Ramagopalan S, Talbot K, Goldacre MJ. Autoimmune disease preceding amyotrophic lateral sclerosis: An epidemiologic study. Neurology 2013;81. https://doi.org/10.1212/WNL.0b013e3182a6cc13.
M.S. L, C.M. V, L.M. M, A.J. de R, C.F. S, P. B, et al. Medical history, lifestyle, family history, and occupational risk factors for follicular lymphoma: The interlymph non-hodgkin lymphoma subtypes project. J Natl Cancer Inst Monogr 2014;2014.
Kwon HJ, Bang DW, Kim EN, Wi C Il, Yawn BP, Wollan PC, et al. Asthma as a risk factor for zoster in adults: A population-based case-control study. Journal of Allergy and Clinical Immunology 2016;137. https://doi.org/10.1016/j.jaci.2015.10.032.
Iribarren C, Tolstykh I V., Miller MK, Sobel E, Eisner MD. Adult asthma and risk of coronary heart disease, cerebrovascular disease, and heart failure: A prospective study of 2 matched cohorts. Am J Epidemiol 2012;176. https://doi.org/10.1093/aje/kws181.
Han YY, Forno E, Celedón JC. Urinary caffeine and caffeine metabolites, asthma, and lung function in a nationwide study of U.S. adults. Journal of Asthma 2022;59. https://doi.org/10.1080/02770903.2021.1993250.
Lynch AM, Patnaik JL, Cathcart JN, Mathias MT, Siringo FS, Lacey Echalier E, et al. Colorado age-related macular degeneration registry: Design and clinical risk factors of the cohort. Retina 2019;39. https://doi.org/10.1097/IAE.0000000000002023.
Söderberg KC, Jonsson F, Winqvist O, Hagmar L, Feychting M. Autoimmune diseases, asthma and risk of haematological malignancies: A nationwide case-control study in Sweden. Eur J Cancer 2006;42. https://doi.org/10.1016/j.ejca.2006.04.021.
Bich TCT, Quoc QL, Choi Y, Yang EM, Trinh HKT, Shin YS, et al. Serum Amyloid A1: A Biomarker for Neutrophilic Airway Inflammation in Adult Asthmatic Patients. Allergy Asthma Immunol Res 2022;14. https://doi.org/10.4168/aair.2022.14.1.40.
Kim TW, Kim MN, Kwon JW, Kim KM, Kim SH, Kim W, et al. Risk of hepatitis B virus reactivation in patients with asthma or chronic obstructive pulmonary disease treated with corticosteroids. Respirology 2010;15. https://doi.org/10.1111/j.1440-1843.2010.01798.x.
Cakir M, Karakas T, Orhan F, Okten A, Gedik Y. Atopy in children with chronic hepatitis B virus infection. Acta Paediatrica, International Journal of Paediatrics 2007;96. https://doi.org/10.1111/j.1651-2227.2007.00405.x.
von Mutius E, Smits HH. Primary prevention of asthma: from risk and protective factors to targeted strategies for prevention. The Lancet 2020;396. https://doi.org/10.1016/S0140-6736(20)31861-4.
Finan C, Gaulton A, Kruger FA, Lumbers RT, Shah T, Engmann J, et al. The druggable genome and support for target identification and validation in drug development. Sci Transl Med 2017;9. https://doi.org/10.1126/scitranslmed.aag1166.
Cotto KC, Wagner AH, Feng YY, Kiwala S, Coffman AC, Spies G, et al. DGIdb 3.0: A redesign and expansion of the drug-gene interaction database. Nucleic Acids Res 2018;46. https://doi.org/10.1093/nar/gkx1143.
Klein R, Knudtson MD, Klein BEK, Wong TY, Cotch MF, Liu K, et al. Inflammation, Complement Factor H, and Age-Related Macular Degeneration. The Multi-Ethnic Study of Atherosclerosis. Ophthalmology 2008;115. https://doi.org/10.1016/j.ophtha.2008.03.021.
Radeke MJ, Radeke CM, Shih YH, Hu J, Bok D, Johnson L V., et al. Restoration of mesenchymal retinal pigmented epithelial cells by TGFβ pathway inhibitors: Implications for age-related macular degeneration. Genome Med 2015;7. https://doi.org/10.1186/s13073-015-0183-x.
Olafsdottir TA, Theodors F, Bjarnadottir K, Bjornsdottir US, Agustsdottir AB, Stefansson OA, et al. Eighty-eight variants highlight the role of T cell regulation and airway remodeling in asthma pathogenesis. Nat Commun 2020;11. https://doi.org/10.1038/s41467-019-14144-8.
Rotin D, Kumar S. Physiological functions of the HECT family of ubiquitin ligases. Nat Rev Mol Cell Biol 2009;10. https://doi.org/10.1038/nrm2690.
Song Y, Wei J, Li R, Fu R, Han P, Wang H, et al. Tyrosine kinase receptor B (TrkB) attenuates liver fibrosis via inhibiting TGF-β/SMAD signaling. Hepatology 2023:10–1097.
Breiteneder H, Peng YQ, Agache I, Diamant Z, Eiwegger T, Fokkens WJ, et al. Biomarkers for diagnosis and prediction of therapy responses in allergic diseases and asthma. Allergy: European Journal of Allergy and Clinical Immunology 2020;75. https://doi.org/10.1111/all.14582.
Cardenas A, Sordillo JE, Rifas-Shiman SL, Chung W, Liang L, Coull BA, et al. The nasal methylome as a biomarker of asthma and airway inflammation in children. Nat Commun 2019;10. https://doi.org/10.1038/s41467-019-11058-3.
Tang R, Yin L, Yao L, Zhang Q-F, Chen X-P. A Co-Expressed Natural Antisense RNA FCER1A-AS Controls IgE-Dependent Immunity by Promoting Expression of FcεRIα. Microbiol Spectr 2023:e00733-23.
Galli SJ, Tsai M, Piliponsky AM. The development of allergic inflammation. Nature 2008;454. https://doi.org/10.1038/nature07204.
Peiper SC, Wang Z xuan, Neote K, Martin AW, Showell HJ, Conklyn MJ, et al. The duffy antigen/receptor for chemoklnes (DARC) is expressed in endothelial cells of duffy negative individuals who lack the erythrocyte receptor. Journal of Experimental Medicine 1995;181. https://doi.org/10.1084/jem.181.4.1311.
Vergara C, Tsai YJ, Grant A V., Rafaels N, Gao L, Hand T, et al. Gene encoding duffy antigen/receptor for chemokines is associated with asthma and IgE in three populations. Am J Respir Crit Care Med 2008;178. https://doi.org/10.1164/rccm.200801-182OC.
Pierron D, Heiske M, Razafindrazaka H, Pereda-Loth V, Sanchez J, Alva O, et al. Strong selection during the last millennium for African ancestry in the admixed population of Madagascar. Nat Commun 2018;9. https://doi.org/10.1038/s41467-018-03342-5.
Efremov IS, Asadullin AR, Dobrodeeva VS, Shnayder NA, Akhmetova EA, Tukhvatullina DR, et al. Association of polymorphic variants of genes (HTR2A, MTNR1A, MTNR1B, CLOCK, DRD2) and insomnia in alcohol dependence syndrome. Nevrologiya, Neiropsikhiatriya, Psikhosomatika 2021;13. https://doi.org/10.14412/2074-2711-2021-5-34-39.
Tran NQV, Le M-K, Nguyen T-A, Kondo T, Nakao A. Association of Circadian Clock Gene Expression with Pediatric/Adolescent Asthma and Its Comorbidities. Int J Mol Sci 2023;24:7477.

Download PDF

Version 1

posted

You are reading this latest preprint version

Discovering causal genes and comorbidities for asthma

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials and methods

2.1. Building the lung and whole blood gene regulatory networks

2.2. Mendelian randomisation analysis

2.3. Biotype annotation of causal asthma genes

2.4. Gene Ontology analysis

2.5. Identification of immune cell types and tissues having shared regulatory patterns of causal asthma genes

2.6. Identification of asthma-associated traits

2.7. DisGeNet and comorbidity analysis

2.8. Ethics

3. Results

3.1. eQTL analysis identifies spatially constrained eQTLs in the lung and blood

3.2. Mendelian randomisation identifies genes causally associated with asthma in the lung and blood

3.3. Seven protein-coding genes within the 17q12-21 locus are causal for asthma

3.5. Asthma endotypes share regulatory mechanisms with the asthma GRNs

3.6. Additional traits are associated with asthma through indirect interactions with the asthma GRNs

3.7. Analysis of gene-disease associations corroborates identified asthma comorbidities observed in New Zealand patients.

3.8. Genes bridging asthma-trait interactions are part of the druggable genome

4. Discussion

Declarations

References

Supplementary Files

Status:

Version 1