1. Epidemiology and Global Health
  2. Genetics and Genomics

A meta-analysis of genome-wide association studies of childhood wheezing phenotypes identifies ANXA1 as a susceptibility locus for persistent wheezing

  1. Raquel Granell  Is a corresponding author
  2. John A Curtin
  3. Sadia Haider
  4. Negusse Tadesse Kitaba
  5. Sara A Mathie
  6. Lisa G Gregory
  7. Laura L Yates
  8. Mauro Tutino
  9. Jenny Hankinson
  10. Mauro Perretti
  11. Judith M Vonk
  12. Hasan S Arshad
  13. Paul Cullinan
  14. Sara Fontanella
  15. Graham C Roberts
  16. Gerard H Koppelman
  17. Angela Simpson
  18. Steve W Turner
  19. Clare S Murray
  20. Clare M Lloyd
  21. John W Holloway
  22. Adnan Custovic  Is a corresponding author
  23. on behalf of UNICORN and Breathing Together investigators
  1. MRC Integrative Epidemiology Unit, Department of Population Health Sciences, Bristol Medical School, University of Bristol, United Kingdom
  2. Division of Infection, Immunity and Respiratory Medicine, School of Biological Sciences, The University of Manchester, Manchester Academic Health Science Centre, and Manchester University NHS Foundation Trust, United Kingdom
  3. National Heart and Lung Institute, Imperial College London, United Kingdom
  4. Human Development and Health, Faculty of Medicine, University of Southampton, United Kingdom
  5. William Harvey Research Institute, Barts and The London School of Medicine Queen Mary University of London, United Kingdom
  6. Department of Epidemiology, University of Groningen, University Medical Center Groningen\, Netherlands
  7. University of Groningen, University Medical Center Groningen, Groningen Research Institute for Asthma and COPD (GRIAC), Netherlands
  8. NIHR Southampton Biomedical Research Centre, University Hospitals Southampton NHS Foundation Trust, United Kingdom
  9. David Hide Asthma and Allergy Research Centre, United Kingdom
  10. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, United Kingdom
  11. Department of Pediatric Pulmonology and Pediatric Allergology, University of Groningen, University Medical Center Groningen, Beatrix Children’s Hospital, Netherlands
  12. Child Health, University of Aberdeen, United Kingdom
https://doi.org/10.7554/eLife.84315
Share this article
Cite this article
  1. Raquel Granell
  2. John A Curtin
  3. Sadia Haider
  4. Negusse Tadesse Kitaba
  5. Sara A Mathie
  6. Lisa G Gregory
  7. Laura L Yates
  8. Mauro Tutino
  9. Jenny Hankinson
  10. Mauro Perretti
  11. Judith M Vonk
  12. Hasan S Arshad
  13. Paul Cullinan
  14. Sara Fontanella
  15. Graham C Roberts
  16. Gerard H Koppelman
  17. Angela Simpson
  18. Steve W Turner
  19. Clare S Murray
  20. Clare M Lloyd
  21. John W Holloway
  22. Adnan Custovic
  23. on behalf of UNICORN and Breathing Together investigators
(2023)
A meta-analysis of genome-wide association studies of childhood wheezing phenotypes identifies ANXA1 as a susceptibility locus for persistent wheezing
eLife 12:e84315.
https://doi.org/10.7554/eLife.84315
  2. Download BibTeX
  3. Download .RIS

Abstract

Background:

Many genes associated with asthma explain only a fraction of its heritability. Most genome-wide association studies (GWASs) used a broad definition of ‘doctor-diagnosed asthma’, thereby diluting genetic signals by not considering asthma heterogeneity. The objective of our study was to identify genetic associates of childhood wheezing phenotypes.

Methods:

We conducted a novel multivariate GWAS meta-analysis of wheezing phenotypes jointly derived using unbiased analysis of data collected from birth to 18 years in 9568 individuals from five UK birth cohorts.

Results:

Forty-four independent SNPs were associated with early-onset persistent, 25 with pre-school remitting, 33 with mid-childhood remitting, and 32 with late-onset wheeze. We identified a novel locus on chr9q21.13 (close to annexin 1 [ANXA1], p<6.7 × 10-9), associated exclusively with early-onset persistent wheeze. We identified rs75260654 as the most likely causative single nucleotide polymorphism (SNP) using Promoter Capture Hi-C loops, and then showed that the risk allele (T) confers a reduction in ANXA1 expression. Finally, in a murine model of house dust mite (HDM)-induced allergic airway disease, we demonstrated that anxa1 protein expression increased and anxa1 mRNA was significantly induced in lung tissue following HDM exposure. Using anxa1-/- deficient mice, we showed that loss of anxa1 results in heightened airway hyperreactivity and Th2 inflammation upon allergen challenge.

Conclusions:

Targeting this pathway in persistent disease may represent an exciting therapeutic prospect.

Funding:

UK Medical Research Council Programme Grant MR/S025340/1 and the Wellcome Trust Strategic Award (108818/15/Z) provided most of the funding for this study.

Editor's evaluation

The study uses a novel meta-analysis approach coupled with endotype discovery in GWAS studies of childhood wheezing to identify ANXA1 as a susceptibility locus. Functional data strengthens the conclusions from a relatively small sample size by GWAS standards. This is representative of a way forward for efficient genetic discovery in deeply phenotyped complex diseases, where there is much unrecognised heterogeneity.

https://doi.org/10.7554/eLife.84315.sa0

eLife digest

Three-quarters of children hospitalized for wheezing or asthma symptoms are preschool-aged. Some will continue to experience breathing difficulties through childhood and adulthood. Others will undergo a complete resolution of their symptoms by the time they reach elementary school. The varied trajectories of young children with wheezing suggest that it is not a single disease. There are likely different genetic or environmental causes.

Despite these differences, wheezing treatments for young children are ‘one size fits all.’ Studying the genetic underpinnings of wheezing may lead to more customized treatment options.

Granell et al. studied the genetic architecture of different patterns of wheezing from infancy to adolescence. To do so, they used machine learning technology to analyze the genomes of 9,568 individuals, who participated in five studies in the United Kingdom from birth to age 18. The experiments found a new genetic variation in the ANXA1 gene linked with persistent wheezing starting in early childhood. By comparing mice with and without this gene, Granell et al. showed that the protein encoded by ANXA1 controls inflammation in the lungs in response to allergens. Animals lacking the protein develop worse lung inflammation after exposure to dust mite allergens.

Identifying a new gene linked to a specific subtype of wheezing might help scientists develop better strategies to diagnose, treat, and prevent asthma. More studies are needed on the role of the protein encoded by ANXA1 in reducing allergen-triggered lung inflammation to determine if this protein or therapies that boost its production may offer relief for chronic lung inflammation.

Introduction

Asthma is a complex disorder caused by a variety of mechanisms which result in multiple clinical phenotypes (Pavord et al., 2018). It has a strong genetic component, and twin studies estimate its heritability to be ~60–70% (Duffy et al., 1990). ‘Asthma genes’ have been identified through a range of approaches, from candidate gene association studies (Simpson et al., 2012) and family-based genome-wide linkage analyses (Daniels et al., 1996) to genome-wide association studies (GWASs) (Moffatt et al., 2007; Moffatt et al., 2010; Demenais et al., 2018). The first asthma GWAS (2007) identified multiple markers on chromosome 17q21 associated with childhood onset asthma (Moffatt et al., 2007). A comprehensive review summarising the results of 42 GWASs of asthma and asthma-related traits has been published recently (El-Husseini et al., 2020). The most widely replicated locus is 17q12-21, followed by 6p21 (HLA region), 2q12 (IL1RL1/IL18R1), 5q22 (TSLP), and 9p24 (IL33) (Kim and Ober, 2019). Overall, the evidence suggests that multiple genes are underlying the association peaks (Kim and Ober, 2019).

However, despite undeniable successes, genetic studies of asthma have produced relatively heterogeneous results, and only a small proportion of the heritability is accounted for (Ober and Yao, 2011). One part of the explanation for the paucity of precise replication are numerous gene-environment interactions (Custovic et al., 2012). Another important consideration is asthma heterogeneity, in that asthma diagnosis comprises several conditions with distinct pathophysiology (Custovic, 2020; Haider et al., 2022), each potentially underpinned by different genetic associations (Custovic et al., 2019). However, in order to maximise sample size, most GWASs used a definition of ‘doctor-diagnosed asthma’ (Aaron et al., 2017). Such aggregated outcome definitions are imprecise (Looijmans-van den Akker et al., 2016) and phenotypically and mechanistically heterogeneous (Robinson et al., 2021), and this heterogeneity may dilute important genetic signals (Custovic et al., 2019).

One way of disaggregating asthma diagnosis is to use data-driven methods to derive subtypes in a hypothesis-neutral way (Howard et al., 2015). For example, we jointly modelled data on wheezing from birth to adolescence in five UK population-based birth cohorts and identified five distinct phenotypes (Oksel et al., 2019a). However, although latent modelling approaches have been instrumental in elucidating the heterogenous nature of childhood asthma diagnosis (Haider et al., 2022), there has been little research into the genetic associations of phenotypes derived using data-driven methods. This is the first study to investigate the genetic architecture of wheezing phenotypes from infancy to adolescence, to identify genes specific to each phenotype and better understand the genetic heterogeneity between the disease class profiles.

Materials and methods

Study design, setting, participants, and data sources/measurement

Request a detailed protocol

The Study Team for Early Life Asthma Research (STELAR) consortium (Custovic et al., 2015) brings together five UK population-based birth cohorts: Avon Longitudinal Study of Parents and Children (ALSPAC) (Golding et al., 2001), Ashford (Cullinan et al., 2004) and Isle of Wight (IOW) (Arshad et al., 2018) cohorts, Manchester Asthma and Allergy Study (MAAS) (Custovic et al., 2002), and the Aberdeen Study of Eczema and Asthma to Observe the Effects of Nutrition (SEATON) (Martindale et al., 2005). All studies were approved by research ethics committees. See Appendix 1: Description of cohorts for more details. Informed consent was obtained from parents, and study subjects gave their assent/consent when applicable.

Validated questionnaires were completed on multiple occasions from infancy to adolescence (Oksel et al., 2019a). A list of variables, per cohort, is shown in Appendix 1—table 1, and the cohort-specific time points and sample sizes in Appendix 1—table 2. Data were harmonised and imported into Asthma eLab web-based knowledge management platform to facilitate joint analyses (Custovic et al., 2015).

Definition of primary outcome (wheeze phenotypes from infancy to adolescence)

Request a detailed protocol

In the pooled analysis among 15,941 subjects with at least two observations on current wheeze, we used latent class analysis (LCA) to derive wheeze phenotypes from birth to age 18 years (Oksel et al., 2019a). A detailed description of the analysis is presented in Oksel et al., 2019a, and in Appendix 1: Definition of variables. A five-class solution was selected as the optimal model (Oksel et al., 2019a), and the classes (wheeze phenotypes) were labeled as: (1) never/infrequent wheeze (52.4%); (2) early-onset pre-school remitting wheeze (18.6%); (3) early-onset middle-childhood remitting wheeze (9.8%); (4) early-onset persistent wheeze (10.4%); and (5) late-onset wheeze (8.8%). These latent classes were used in the subsequent GWAS.

Genotyping, imputation, and GWAS meta-analysis

Request a detailed protocol

Genotyping, quality control, imputation, and exclusions are described in Appendix 2: Genotyping and imputation. Analyses were performed independently in ALSPAC, MAAS, and the combined IOW-SEATON-Ashford (genotyped on the same platform, at the same time, and imputed together). We used SNPTEST v2.5.2 (Marchini and Howie, 2010) with a frequentist additive multinomial logistic regression model (-method newml), using the never/infrequent wheeze as the reference and without including any covariates. A meta-analysis of the three GWASs was performed using METAL (Willer et al., 2010) with a total of 8,057,852 SNPs. See Appendix 2: LD clumping, pre-selection, and gene annotation for more details.

Post-GWASs

Request a detailed protocol

Our GWAS identified a novel locus in chr9q21 nearby Annexin A1 (ANXA1), exclusively associated with early-onset persistent wheeze (see Results section). We therefore proceeded with studies to identify causal variants and explore the biological mechanisms underlying this locus (see Appendix 3: Post-GWAS: rs75260654 (ANXA1) for more details). To this end, we firstly identified the most likely causative SNP using Promoter Capture Hi-C (PCHi-C) loops. We then ascertained genotype effect on gene expression and assessed the potential biological function of ANXA1 in asthma. Finally, we used a murine model of house dust mite (HDM)-induced allergic airway disease to investigate whether ANXA1 was important in regulating immune responses to a clinically relevant aeroallergen and used knock-out mice to derive further in vivo functional data to support our GWAS finding.

Results

Participants and descriptive data

We included a total of 9568 subjects with European ancestry: ALSPAC, n=6833; MAAS, n=887; SEATON, n=548; Ashford, n=348; and IOW, n=952. Demographic characteristics of the participants in STELAR cohorts included in this analysis and a flowchart are shown in Appendix 1—table 3 and Appendix 1—figure 1. Cohorts contain similar proportions of males (range 48–54%), maternal history of asthma (11–14%), maternal smoking (14–23%), (doctor-diagnosed) asthma ever during mid-childhood (16–24%) and adolescence (20–30%), current wheeze (12–20% mid-childhood, 9–25% adolescence), and current use of asthma medication (12–17% mid-childhood, 11–17% adolescence). Individuals with missing genetic data as well as related and non-European individuals were excluded. Comparison of included vs. excluded individuals across cohorts (per cohort and time point) is in Appendix 1 and Appendix 1—table 4.

GWAS meta-analysis

We conducted three GWASs (ALSPAC, MAAS, IOW-SEATON-Ashford) in parallel and results were meta-analysed. The distribution of the minor allele frequencies was consistent across genotyped datasets (mean SD 0.01). A circular Manhattan plot and a QQ plot are shown in Figure 1, Figure 1—figure supplement 1. Some observed p-values were clearly more significant than expected under the null hypothesis, particularly for early-onset persistent wheeze, without an early separation of the expected from the observed which indicates low evidence of population stratification. We observed slight deflation of the meta-analysis p-values in our summary statistics. Genomic inflation factor (λ) for early-onset pre-school remitting = 0.96, early-onset mid-childhood remitting = 0.94, late-onset = 0.96, and early-onset persistent wheezing = 0.97. A total of 589 SNPs were associated with at least one phenotype with p<10–5. After clumping, we identified 134 independent SNPs uniquely associated with different phenotypes (p<10–5): of these, 44 were exclusively associated with early-onset persistent, 25 with early-onset pre-school remitting, 33 with early-onset mid-childhood remitting, and 32 with late-onset wheeze (Appendix 2—table 1). Scatter plots in Figure 2, Figure 2—figure supplement 1 show the heterogeneity in the genetic profile of the wheeze phenotypes. The plots show that all signals were phenotype-specific at p<10–5 and only nominal associations were shared across wheezing phenotypes. More details on how these plots were derived can be found in Appendix 2: Heterogeneity scatter plots. For example, chr17q21 was identified as a top locus for early-onset persistent wheeze (p=5.42 × 10–9), but some of the SNPs in this region were also associated with the early-onset mid-childhood remitting phenotype (p<10–4).

Figure 1 with 1 supplement see all
Download asset Open asset
Circular Manhattan plot showing an overview of the genome-wide association study (GWAS) results by wheeze phenotype (from outside to inside: early-onset persistent, early-onset pre-school remitting, early-onset mid-childhood remitting, and late-onset wheeze).

The red line indicates the genome-wide significance threshold (p < 5 × 10−8), while the blue line indicates the threshold for genetic variants that showed a suggestive significant association (p < 10−5).

Figure 2 with 1 supplement see all
Download asset Open asset
Scatter plots illustrating the heterogeneity in the genetic profile of the wheezing phenotypes.

Top plots compare phenotype-specific beta effects for persistent and early-onset mid-childhood remitting wheezing. Shared nominal beta effects only found when relaxing p<10–4 for early-onset mid-childhood remitting wheezing. Bottom plots compare phenotype-specific beta effects for persistent and early-onset pre-school remitting. No shared beta effects (same or opposite direction) were found at p<10–5 for any of the comparisons. Abbreviations used: pw = persistent, er = early-onset pre-school remitting, and pe = early-onset mid-childhood remitting.

To help identify functional elements located near the GWAS-associated variants (potential causal variants), we used locus zoom plots (LZPs) for the 134 independent SNPs (p<10–5). Following close inspection of all plots, we short-listed 85 independent SNPs (Appendix 2—table 1) for which the LZPs potentially indicated more than one causal variant (Appendix 2—figures 1–4) and followed them up for further annotation. The results of GWAS meta-analysis for these 85 SNPs with main associations across the four wheeze phenotypes are presented in Table 1. Previously associated traits for each region/gene associated with the different wheeze phenotypes are shown in Appendix 4—tables 1–4 and results are summarised in Appendix 4: Results in context of literature. Briefly, one region (6q27) among the top hits for early-onset pre-school remitting wheeze was previously associated with asthma, but in the context of obesity with a nominal association with asthma and BMI (Melén et al., 2010). Another region/gene (3q26.31/NAALADL2) identified as top hit for early-onset pre-school remitting wheeze was reported as an associate of severe asthma exacerbations, but only at nominal level (Herrera‐Luis et al., 2021). No regions/genes identified as top hits for early-onset mid-childhood remitting wheeze were found to have previous associations with asthma. Several genes/loci identified as top hits for late-onset wheeze were previously associated with asthma: ACOXL chr2q13 (later onset asthma and obesity; Zhu et al., 2020), PRKAA2 chr1p32.2 (lymphocyte count and asthma susceptibility; Cusanovich et al., 2012), CD200 3q13.2 (adult onset non-allergic asthma; Siroux et al., 2014), GIMAP family 7q36.1 (autoimmune diabetes, asthma, allergy; Heinonen et al., 2015), 9p22.3 (asthma in <16 years of age; Denham et al., 2008), and 16p12.1 (asthma and rhino-conjunctivitis at 10–15 years; Sottile et al., 2019).

Table 1
Genome-wide association study (GWAS) meta-analysis: short-listed 85 top independent single nucleotide polymorphisms (SNPs) across the four wheezing phenotypes.
Early-onset persistent wheezing
LocusIndependent SNPsNearby genes (SNPnexus)Effect allele(freq)/ other alleleBetaSEp-ValueEffect direction (3 GWAS)min_pval_other*Previous relevant associations†
1q43rs4620530CHRM3g(0.56)/t0.250.052.45E-06+++0.79FEV1, FEV1/FVC, asthma- high priority drug target
2p25.1rs13398488RNF144Ag(0.29)/a0.250.052.18E-06+--0.13Asthma, allergy, childhood onset asthma, allergic rhinitis
2q12.2rs6543291FHL2c(0.4)/t0.230.056.97E-06+++0.10Bronchial hyper-responsiveness, airway inflammation; novel gene associated with asthma severity in human
3q21.3rs77655717EFCC1, RAB43, RAB7Ac(0.05)/t0.470.106.40E-06+++0.39RAB43: response to bronchodilator, FEV/FEC ratio; RAB7A: eosinophil count
4p16.3rs7680608eQTLRNF212, IDUA, DGKQ, SLC26A1g(0.93)/c–0.420.091.31E-06---0.154p16: asthma
rs77822621eQTLc(0.96)/t–0.500.117.16E-06---0.01
4q31.21rs115228498INPP4Bc(0.02)/t0.790.172.70E-06+++0.02Atopic asthma
5p15.31rs116494115ADCY2g(0.01)/a0.750.176.49E-06+++0.09Asthma×air pollution, childhood asthma
7q22.3rs76871421CDHR3c(0.12)/t0.370.075.71E-07+++0.22Childhood asthma
9q21.13rs75260654ANXA1, TMC1, LOC101927258, ALDH1A1c(0.98)/t–0.900.166.66E-09---0.05ANXA1: FEV1/FVC, response to bronchodilators in smokers, with anti-inflammatory properties, strongly expressed in bronchial mast cells and potentially involved in epithelial airway repair
rs116849664c(0.98)/t–0.890.161.99E-08---0.06
10q24.2rs7088157LOXL4, R3HCC1Lg(0.5)/a–0.230.057.34E-06---0.26R3HCC1L: eosinophil count, atopic eczema, psoriasis, BMI
11p15.4rs112474574TRIM5, TRIM6, TRIM22c(0.96)/t–0.550.122.29E-06---0.14Severe asthma and insulin resistance
11q23.3rs116861530eQTLSIK3g(0.94)/a–0.420.099.07E-06---0.01Triglycerides, glucose metabolism, eosinophil count
14q22.1rs1105683KTN1c(0.07)/t0.410.099.15E-06+++0.24Severe asthma
5q13.3rs2202714eQTLFAM227Bg(0.36)/a0.230.058.71E-06+++0.01rs35251997 and FEV1; FEV1/FVC
15q25.2rs117540214eQTLADAMTSL3g(0.06)/a0.420.109.82E-06+++3.91E-03FEV1/FVC
17q12rs17676191IKZF3g(0.10)/a0.360.082.18E-06+++3.06E-03Early-onset asthma, persistent wheezing (chr17q12-q21)
rs79026872c(0.03)/t0.640.132.08E-06+++2.56E-03
17q21rs4795400GSDMBc(0.53)/t0.300.055.42E-09+++1.96E-04
rs1031460g(0.50)/t0.270.058.71E-08+++1.87E-04
rs56199421c(0.45)/t–0.230.054.50E-06---9.61E-04
rs4795406LRRC3Cg(0.55)/c–0.240.059.91E-07---1.51E-03
rs72832972GSDMAc(0.92)/t–0.380.088.91E-06---0.01
rs4794821c(0.47)/t0.270.059.43E-08+++1.07E-03
rs59843584c(0.78)/a–0.310.066.38E-08---6.63E-03
19p13.2rs4804311MARCH2, HNRNPM, MYO1Fg(0.08)/a0.420.099.65E-07+-+0.05Triglycerides, HDL cholesterol, metabolic syndrome; MYO1F: FEV1 and FVC
rs2013694c(0.89)/t–0.380.088.29E-07--+0.39
rs73501545g(0.16)/a0.310.078.39E-06+++0.29
rs111644945g(0.9)/a–0.410.084.01E-07---0.02
22q11.1rs5994170CECR5g(0.4)/a0.230.054.95E-06+++0.58Triglycerides, eosinophil count, and body height
rs34902370c(0.75)/t–0.250.066.80E-06---0.41
Early-onset pre-school remitting wheezing
LocusSNPNearby genes (SNPnexus)Coded(freq)/other alleleBetaSEp-ValueDirectionmin_pval_otherPrevious relevant associations
1q32.3rs12730098eQTLPPP2R5Ac(0.79)/t–0.220.058.44E-06---0.53Waist circumference and obesity
2p24.2rs2880066FAM49A or CYRIAt(0.09)/a0.320.074.34E-06+++0.20Airway repair in non-atopic asthma
rs10180268c(0.06)/t0.430.096.56E-07+++0.19
3q26.31rs3861377NLGN1g(0.89)/a–0.280.067.75E-06---0.28Smoking
rs10513743NAALADL2c(0.84)/t–0.250.064.97E-06-+-0.06Exacerbations requiring hospitalisation in asthma-suggestive p-value
5q13.3rs10075253SV2Cc(0.85)/t–0.270.061.20E-06---0.17BMI
6q27rs2453395PDE10Ag(0.33)/a0.190.049.51E-06+++0.01Birth weight; asthma and BMI
7q21.11rs4730561MAGI2g(0.36)/a–0.200.046.78E-06---0.13Allergic diseases and atopy, smoking, BMI, airway wall thickness
rs73144976g(0.97)/a–0.470.119.41E-06---0.26
rs67259321c(0.06)/t0.430.081.65E-07+-+0.76
9p13.3rs10758259eQTLC9orf24g(0.17)/a–0.270.064.64E-06---0.01Airway repair
11q22.3rs72994149GUCY1A2c(0.84)/t–0.240.058.33E-06-+-0.06Systolic blood pressure
13q21.1rs2872948PRR20A/B/C/D/Et(0.96)/a–0.540.105.93E-08---0.27Systolic blood pressure
rs73527654g(0.08)/a0.340.072.85E-06+++0.41
15q21.1rs116966886SEMA6Dg(0.99)/a–0.820.187.55E-06-+-0.57Smoking
rs117565527g(0.99)/a–0.870.172.38E-07-+-0.43
Early-onset mid-childhood remitting wheezing
LocusSNPNearby genes (SNPnexus)Coded(freq)/ other alleleBetaSEp-ValueDirectionmin_pval_otherPrevious relevant associations
1q23.2rs35725789CADM3, FCER1A, MPTX1, OR10J1c(0.95)/a–0.560.125.42E-06-+-0.01Neutrophil count, CRP
rs146141555c(0.98)/t–0.890.172.04E-07-+-0.08
rs146575092g(0.98)/a–0.850.178.73E-07-+-0.07
2p22.3rs7595553MRPL50P1g(0.16)/c–0.460.103.26E-06---0.12PM 2.5 exposure level and global DNA methylation level
3p25.3rs34315999eQTLRAD18c(0.03)/t0.690.141.11E-06+++0.14Atopy/SPT
3q29rs146961758MRPL50P1, LSG1, TMEM44-AS1, TMEM44, ATP13A3t(0.05)/a0.570.126.01E-06+-+0.113q29: BMI
TMEM44-AS1, TMEM44, ATP13A3: diastolic blood pressure; LSG1: BMI, eosinophil count
4q24rs138794367SLC9B1c(0.99)/t–1.020.225.47E-06---0.13Eosinophil count, allergic rhinitis
5q14.1rs115719402AP3B1g(0.96)/a–0.600.137.20E-06---0.06Vital capacity, BMI
13q31.1rs9602218RNU6-67P, SLITRK1c(0.06)/a0.580.121.74E-06+-+0.05RNU6-67P/ rs976078: food allergy
rs61960366g(0.97)/a–0.790.157.09E-08-+-0.12
rs74589927g(0.02)/a0.730.163.78E-06+-+0.02
rs2210726VENTXP2, UBE2D3P4, MTND4P1c(0.91)/t–0.470.101.33E-06---0.02
rs4390476c(0.08)/a0.460.108.81E-06+++0.12
14q24.2rs117443464ZFYVE1g(0.95)/a–0.570.124.68E-06--+0.19LDL cholesterol and systolic blood pressure
20p12.3-p12.2rs6077514PLCB4c(0.88)/t–0.390.094.03E-06---0.43Neutrophil count
Late-onset wheezing
LocusSNPNearby genes (SNPnexus)Coded(freq)/ other alleleBetaSEp-ValueDirectionmin_pval_otherPrevious relevant associations
1p36.13rs9439669KLHDC7At(0.82)/a–0.340.075.15E-06---0.311p36.13: metabolic syndrome
1p32.2rs2051039PPAP2B, PRKAA2c(0.08)/t0.470.106.06E-06+-+0.08PRKAA2: lymphocyte count and asthma susceptibility
1p31.1rs72673642HMGB1P18g(0.77)/a–0.310.076.25E-06---0.01Smoking, BMI
2q13rs140983998ACOXL, BUB1c(0.98)/t–0.880.194.71E-06---0.40ACOXL: later onset asthma and obesity
2q14.3rs148008098AMMECR1Lc(0.96)/t–0.690.153.41E-06---0.01Body height, blood protein; growth, bone, and heart alterations
3p24.2rs4072729RARBc(0.03)/t0.610.134.20E-06+-+0.23FEV1/FVC, adult lung function
3q13.2rs145629570KIAA2018, NAA50, SIDT1, CD200c(0.02)/t0.920.186.83E-07+++0.10SIDT1: FEV1/FVC; CD200: adult-onset non-allergic asthma
3q23rs113643470TFDP2, XRN1c(0.98)/t–0.910.191.68E-06---0.03XRN1: eosinophil count; 3q23: allergic disease and atopic sensitisation
4p11rs17472015SLAIN2, SLC10A4, FRYLc(0.01)/t1.000.239.49E-06+++0.46FRYL: body height, age at menopause
7q36.1rs117660982KRBA1, ZNF467g(0.97)/a–0.740.167.63E-06-+-0.18Systolic blood pressure
rs118027705GIMAP family, AOC1c(0.97)/t–0.770.176.48E-06-+-0.01AOC1: CV disease, smoking; GIMAP family: autoimmune diabetes, asthma, and allergy
rs139489493LOC105375566c(0.98)/t–0.950.202.28E-06---0.03
7q36.3rs144271668PTPRN2c(0.01)/a0.880.192.91E-06+++0.28Eczema
8q21.3rs990182LOC105375631t(0.42)/a0.280.062.57E-06+++0.468q21.3: type 1 diabetes
9p22.3rs79110962NFIB, ZDHHC21c(0.08)/t0.510.103.98E-07+++0.059p22.3: asthma (mean age <16 years)
10q23.31rs7896106SLC16A12, IFIT family, PANK1g(0.35)/t0.300.061.35E-06+++0.05SLC16A12: Body height; PANK1: insulin
11q23.3rs141958628CBL, CCDC84, MCAMc(0.98)/t–0.980.201.33E-06-+-0.27CCDC84: asthma, allergy
15q15.3-q21.1rs139134265SPG11, CTDSPL2g(0.02)/c0.870.209.11E-06+-+0.13CTDSPL2: alcohol drinking
15q25.2rs143862030ADAMTSL3, GOLGA6L4, UBE2Q2P8c(0.04)/t0.640.131.65E-06+++0.08ADAMTSL3: FEV1/FVC; lean mass
16p13.3rs113390367SSTR5-AS1, CACNA1Hg(0.86)/a–0.400.081.04E-06---0.16CACNA1H: eosinophil count
16p12.1rs4788025GSG1Lg(0.46)/a–0.300.067.99E-07---0.1916p12.1: current asthma and rhino-conjunctivitis at 10–15 years
22q13.32rs133498FAM19A5 or TAFA5g(0.94)/a–0.480.115.35E-06---0.84Obesity and metabolic dysfunction

  1. eQTL: identified in expression analyses of whole blood and/or lung tissues using Genotype-Tissue Expression database (https://gtexportal.org) using the European reference panel.

  2. Bold p-values are genome-wide significant (p < 5 × 10−8).

  3. *

    Minimum p-value across associations with the other three wheezing phenotypes, using the never/infrequent wheeze as the baseline phenotype.

  4. †

    List of references or sources (GeneCards, GWAS Catalog, PhenoScanner) available in Appendix 5—tables 1–4.

We identified two GWAS-significant loci for early-onset persistent wheeze: 17q21, p<5.5 × 10–9, and a novel locus on 9q21.13 (ANXA1), p<6.7 × 10–9. The ANXA1 locus was the only GWAS-significant locus that had not previously been associated with asthma or atopic traits, with one previous study showing an association with FEV1/FVC and bronchodilator response in smokers (Lutz et al., 2015). ANXA1 is strongly expressed in bronchial mast cells and has anti-inflammatory properties (Vieira Braga et al., 2019), and may be involved in epithelial airway repair (Leoni et al., 2015; Appendix 4—table 1). We therefore followed up top SNPs from this locus.

ANXA1 locus and persistent wheeze

Two SNPs (rs75260654, the lead SNP, and rs116849664 located downstream of ANXA1) were associated with early-onset persistent wheeze at genome-wide significance (GWS), with an additional SNP rs78320984 almost reaching GWS (Appendix 5—table 1). These SNPs are in linkage disequilibrium (LD) with each other (Appendix 5—figure 1), but not with any other SNPs.

Promoter Capture identifies rs75260654 as the most likely causative variant

To identify the most likely causative variant, we investigated the overlap of the SNPs with PCHi-C interactions involving the ANXA1 promoter in CD4+ cells in MAAS cohort subjects. Of the three SNPs, only rs75260654 overlapped a region interacting with the ANXA1 promoter (Figure 3). Moreover, rs75260654 overlapped a POLR2A ChIP-seq peaks and an ATAC-seq peak and active enhancer in the type II pneumocyte-derived A549 cell line. This shows that rs75260654 is located in a region directly interacting with the ANXA1 promoter and is transcriptionally active in relevant cell types.

Figure 3
Download asset Open asset
Chromatin interactions between rs75260654 and the ANXA1 promoter in CD4+ cells in Manchester Asthma and Allergy Study (MAAS) rs75260654 physically interacts with ANXA1 promoter in CD4+ T cells and overlaps a region of active (POLR2AphosphoS2 ChIP-seq) open (ATAC-seq) chromatin in A549 cell line (lung epithelial carcinoma).

The region is also predicted to be an active enhancer (ChromHMM 18-state model) in the A549 cell type. Only ChromHMM enhancer chromatin are displayed. Yellow shaded area indicates the Promoter Capture Hi-C (PCHi-C) fragment overlapping rs75260654 (red bar) and interacting with the ANXA1 promoter.

Allele frequencies of rs75260654 (MAF = 0.02) across wheeze phenotypes are shown in Appendix 5—table 2. Two individuals (one in MAAS and one in ALSPAC) were homozygote for the minor allele (T), and both were in the early-onset persistent wheeze class. One subject reported current wheeze and asthma through childhood, with hospitalisations for lower respiratory tract infection in the first year of life confirmed in healthcare records. The second individual reported current wheezing at 1.5, 2.5, and 8–9 years and doctor-diagnosed asthma and the use of asthma medication at 8–9 years.

Rs75260654: effect on genomic features

Variant Effect Predictor (VEP) prediction shows the SNP rs75260654 (C changed to T) to be located downstream of three protein-coding transcripts of AXNA1 and overlapping the known regulatory region ID ENSR00000882742 on Chromosome 9: 73,173,001–73,173,200. This region is active in the GI tract, M2 macrophages, neural progenitor cells, and trophoblasts, but is repressed in T lymphocytes including CD4+ CD25+, Treg, and CD8+ cells.

Rs75260654: effect on gene expression

The effect of rs75260654 on the expression of nearby genes was investigated by browsing the eQTL GTEX data available in Ensembl. Compared to C, the T allele was found to reduce the expression of ANXA1 in naïve B cells (effect size = −2.36795, p=0.01) and to increase expression in lymphoblasoid cell lines (effect size = 0.848856, pe = 0.001) (Figure 4). This SNP affects expression of the neighbouring gene ALDH1A1 (aldehyde dehydrogenase-1 family member A1) (effect size = −2.40446, p=0.0039 in macrophages infected with Salmonella). In the eQTL catalogue, rs75260654 is identified as an eQTL of ANXA1 in various immune cells (at nominal significance) including T cells, monocytes, fibroblasts, whole blood, Th2 memory cells, naïve B cells. rs75260654 is also an eQTL of ANXA1 in monocytes that were stimulated with R848 (agonist of TLRs 7 and 8) and a human seasonal influenza A virus (Quach et al., 2016) (at nominal significance) (Appendix 5—table 3). In the lung rs116849664 and rs78320984 (both in LD with rs75260654) were eQTLs of ANXA1 (Appendix 5—table 4) as well as LINC01474 at nominal significance levels.

Figure 4
Download asset Open asset
eQTL ANXA1 and rs75260654 across different tissue types.

Point size is proportional to -log10 p-value.

Additional supporting evidence regarding the significance of the T-allele on the expression of these genes was provided using eQTLGene Consortium meta-analysis of 24 cohorts and 24,331 samples (Võsa et al., 2018). This method reproduced the previous modest results showing a cis-eQTL effect of rs75260654 on both the ANXA1 (p=6.02 × 10–23) and ALDH1A1 (p=1.11 × 10–19) at FDR = 0. No significant trans-eQTLs were observed.

Potential biological function of ANXA1 in asthma

Protein-protein network analysis demonstrated that ANXA1 interacts directly with genes enriched for asthma (including IL4 and IL13) and inflammatory regulation (NR3C1, glucocorticoid receptor) showing its significance in dysregulation of the immune response (see Appendix 5—figure 2 and Appendix 5—table 5).

Functional studies of anxa1 in a murine model

Pulmonary expression of anxa1 is modulated by aeroallergen exposure

We first analysed expression of anxa1 using a model of HDM-induced allergic airway disease (Figure 5A; Gregory et al., 2009). Consistently, immunohistochemistry analysis revealed anxa1 protein expression increased following HDM challenge (Figure 5B and C). Anxa1 mRNA was significantly induced in lung tissue following HDM exposure (Figure 5D). This increase suggests that the pro-resolving anxa1 may play a role in regulating the pulmonary immune response to allergen.

Figure 5
Download asset Open asset
Annexin A1 is induced following house dust mite (HDM) challenge and mice deficient in ANXA1 have exacerbated airway hyper-reactivity.

(A) Schematic of HDM allergen dosing protocol, N=4–6 per group, data representative of two animal experiments. (B, C) Immunofluorescent staining of paraffin-embedded lung tissue sections incubated with anti-annexin A1, counterstained with DAPI (N=4 per group). (D) mRNA expression of annexin A1 in lung tissue following HDM exposure, expression normalised to housekeeping gene hprt. Mice receiving HDM were analysed for changes in airway hyper-reactivity following methacholine (MCh) challenge in tracheotomised restrained mice. (E) Airway resistance at top MCh dose 100 mg/ml. (F) Eosinophils quantified in BAL, (F) T1/ST2+ lymphocytes quantified in the BAL. (H) IL-4, (I) IL-5, and (J) IL-13 quantified in lung tissue by ELISA. *p<0.05 and **p<0.01 relative to PBS control group by Mann-Whitney test. +p < 0.05 and ++p < 0.01 comparing HDM annexin A1 knock-out (KO) mice relative to HDM wildtype (WT) group by Mann-Whitney test.

Anxa1 suppresses allergen-induced airway hyperresponsiveness and type 2 inflammation

To confirm a functional role for anxa1 in allergic airway disease, we exposed anxa1-/- mice to intranasal HDM. Wildtype (WT) mice given HDM over 3 weeks developed significant airway hyperresponsiveness (AHR) compared to PBS control mice. Mice deficient in anxa1 had significantly worse lung function (greater airway resistance) compared to WT-treated mice (Figure 5E). Anxa1-/- mice exhibited significantly increased airway eosinophilia and elevated numbers of Th2 lymphocytes (Figure 5F and G). Lung tissue cytokine levels reflected the exacerbated airway Th2 inflammation, with elevation in IL-4, and significant induction of IL-5 and IL-13 (Figure 5H and J). Thus, anxa1 deficiency results in an alteration of the pulmonary immune response, with uncontrolled eosinophilia and an exacerbation of type 2 inflammation and AHR in response to allergen. More details in Appendix 6: Functional mouse experiments.

Discussion

Herein, we present a comprehensive description of the genetic architecture of childhood wheezing disorders. Using a novel approach applied to a unique dataset from five UK birth cohorts, we identified subsets of SNPs differentially associated across four wheezing phenotypes: early-onset persistent (44 SNPs, 19 loci), early-onset pre-school remitting (25 SNPs, 10 loci), early-onset mid-childhood remitting (33 SNPs, 9 loci), and late-onset (32 SNPs, 20 loci). We found little evidence of genetic associations spanning across different phenotypes. This suggests that genetic architecture of different wheeze phenotypes comprises a limited number of variants likely underpinning mechanisms which are shared across phenotypes, but that each phenotype is also characterised by unique phenotype-specific genetic associations. Importantly, we identified a novel locus in chr9q21 nearby ANXA1 exclusively associated with early-onset persistent wheeze (p<6.7 × 10–9). To identify the most likely causative variant, we investigated the overlap of the associated SNPs with PCHi-C interactions to demonstrate that SNP rs75260654 overlapped a region interacting with the ANXA1 promoter. Using eQTL data, we identified that the risk allele (T) of rs75260654 associated with early-onset persistent wheeze is also associated with ANXA1 expression. Further investigation of the biological function of ANXA1 revealed that it interacts with genes enriched for asthma (including IL4 and IL13) and inflammatory regulation (NR3C1, glucocorticoid receptor). In functional mouse experiments, anxa1 protein expression increased and anxa1 mRNA was significantly induced in lung tissue following HDM exposure, suggesting that the pro-resolving anxa1 may play a role in regulating the pulmonary immune response to allergen. Concurrently, by utilising anxa1-/- deficient mice we demonstrated that loss of anxa1 results in heightened AHR and Th2 inflammation upon allergen challenge, providing important in vivo functional data to support our GWAS finding.

ANXA1 is a 37 kDa glycoprotein with potent anti-inflammatory and pro-resolving properties that are mediated by interaction with a specific G protein-coupled receptor FPR2 (Perretti et al., 2002). This axis represents an important resolution pathway in chronic inflammatory settings such as those of rheumatoid arthritis (D’Acquisto et al., 2008) and ulcerative colitis (Vong et al., 2012). ANXA1 belongs to the annexin family of Ca2+-dependent phospholipid-binding proteins, and through inhibition of phospholipase A2, it reduces eicosanoid production, which also contributes to its anti-inflammatory activities. Modulation of M2 macrophage phenotype is also promoted by ANXA1 to attenuate tissue inflammation (McArthur et al., 2020). Corticosteroids (a mainstay of asthma treatment) increase the synthesis of ANXA1 (Rhen and Cidlowski, 2005). Plasma ANXA1 levels are significantly lower in asthmatic patients with frequent exacerbations compared to those with stable disease, suggesting a link between this mediator and disease state (Lee et al., 2018). Furthermore, children with wheeze have reduced airway levels of ANXA1 ( Eke Gungor et al., 2014).

Previous functional studies using anxa1-/- deficient mice challenged with ovalbumin showed anxa1-deficient mice to have elevated AHR compared to WT mice (Ng et al., 2011). Ng et al. demonstrated that untreated anxa1-deficient mice have spontaneous AHR that predisposes them to exacerbated response to allergen (Ng et al., 2011). In the current study, we demonstrated in the murine lung the induction of Anxa1 in response to HDM exposure. In addition, genetic deletion of anxa1 potentiated the development of AHR and enhanced eosinophilia and markers of Th2 inflammation in mice treated with HDM, which is consistent with and extends previous reports. Of interest, in mice, anxa1 expression was recently found to be characteristic of a novel cell type called the Hillock cell, which may be involved in squamous barrier function and immunomodulation (Montoro et al., 2018). These data identify the ANXA1/FPR2 signalling axis as an important regulator of allergic disease, that could be manipulated for therapeutic benefit.

Our study has several limitations. By GWAS standards, our study is comparatively small and may be considered to be underpowered. The sample size may be an issue when using an aggregated definition (such as ‘doctor-diagnosed asthma’) but is less likely to be an issue when primary outcome is determined by deep phenotyping. This is indirectly confirmed in our analyses. Our primary outcome was derived through careful phenotyping over a period of more than two decades in five independent birth cohorts, and although comparatively smaller than some asthma GWASs, our study proved to be powered enough to detect previously identified key associations (e.g., chr17q21 locus). Precise phenotyping has the potential to identify new risk loci. For example, a comparatively small GWAS (1173 cases and 2522 controls) which used a specific subtype of early-onset childhood asthma with recurrent severe exacerbations as an outcome identified a functional variant in a novel susceptibility gene CDHR3 (SNP rs6967330) as an associate of this disease subtype, but not of doctor-diagnosed asthma (Bønnelykke et al., 2014). This important discovery was made with a considerably smaller sample size but using a more precise asthma subtype. In contrast, the largest asthma GWAS to date had an ~40-fold higher sample size (Demenais et al., 2018), but reported no significant association between CDHR3 and aggregated asthma diagnosis. Therefore, with careful phenotyping, smaller sample sizes may be adequately powered to identify larger effect sizes than those in large GWASs with broader outcome definitions (Schoettler et al., 2019).

The importance of the precise outcome definition was highlighted in our previous studies in ALSPAC which explored genetic associates of wheeze phenotypes derived by LCA (Granell et al., 2013; Spycher et al., 2012). Our current findings are consistent with our earlier report suggesting that 17q21 SNPs are associated with early-onset persistent, but not with early transient or late-onset wheeze (Granell et al., 2013). Further analysis using genetic prediction scores based on 10–200,000 SNPs ranked according to their associations with physician-diagnosed asthma found that the 46 highest ranked SNPs predicted persistent wheeze more strongly than doctor-diagnosed asthma (Spycher et al., 2012). Finally, a candidate gene study combining data from ALSPAC and PIAMA found different associations of IL33-IL1RL1 pathway polymorphisms with different phenotypes (Savenije et al., 2014).

We are cognisant that there may be a perception of the lack of replication of our GWAS findings. We would argue that direct replication is almost certainly not possible in other cohorts, as phenotypes for replication studies should be homogenous (Crawford et al., 2015). However, there is a considerable heterogeneity in LCA-derived wheeze phenotypes between studies, and although phenotypes in different studies are usually designated with the same names, they differ between studies in temporal trajectories, distributions within a population, and associated risk factors (Oksel et al., 2018). This heterogeneity is in part consequent on the number and the non-uniformity of the time points used, and is likely one of the factors responsible for the lack of consistent associations of discovered phenotypes with risk factors reported in previous studies (Oksel et al., 2019b). This will also adversely impact the ability to identify phenotype-specific genetic associates. For example, we have previously shown that less distinct wheeze phenotypes in PIAMA were identified compared to those derived in ALSPAC (Savenije et al., 2011). Thus, phenotypes that are homogeneous to those in our study almost certainly cannot readily be derived in available populations. This is exemplified in our attempted replication of ANXA1 findings in PIAMA cohort (see Appendix 5: Replication of ANXA1 top hits in PIAMA cohort and Appendix 5—table 6). In this analysis, the number of individuals assigned to persistent wheezing in PIAMA was small (Võsa et al., 2018), associates of this phenotype differed to those in STELAR cohorts, and the SNPs’ imputation scores were low (<0.60), which meant the conditions for replication were not met.

Our study population is of European descent, and we cannot generalise the results to different ethnicities or environments. It is important to highlight the under-representation of ethnically diverse populations in most GWASs (Kim and Ober, 2019). To mitigate against this, large consortia have been formed, which combine the results of multiple ethnically diverse GWASs to increase the overall power to identify asthma susceptibility loci. Examples include the GABRIEL (Moffatt et al., 2010), EVE (Torgerson et al., 2011), and TAGC (Demenais et al., 2018) consortia, and the value of diverse, multiethnic participants in large-scale genomic studies has recently been shown (Wojcik et al., 2019). However, such consortia do not have the depth of longitudinal data to allow the type of analyses which we carried out to derive a multivariable primary outcome. Finally, the manual and visual inspection of LZPs for the refinement of association signals and identification of functional elements was an objective approach which might have undermined the findings. One strength of our approach is that we used data from five birth cohorts with detailed and lifelong phenotyping, which were harmonised in a common knowledge management platform (Custovic et al., 2015), allowing joint analyses. We performed three parallel GWASs that produced estimates with remarkably consistent directions of effects.

In conclusion, using unique data from five UK birth cohorts, we identified subsets of SNPs differentially associated across four wheezing phenotypes from infancy to adolescence. We found little evidence of genetic associations spanning across different phenotypes. We discovered a novel locus in chr9q21 uniquely associated with early-onset persistent wheeze (p<6.7 × 10–9), identified SNP rs75260654 as the most likely causative variant, and demonstrated that the risk allele (T) confers a reduction in ANXA1 expression. In mouse experiments, ANXA1 expression increased in lung tissue following allergen exposure, suggesting that the pro-resolving ANXA1 may play a role in regulating the pulmonary immune response to allergen. Using ANXA1-deficient mice, we demonstrated that loss of ANXA1 results in heightened AHR and Th2 inflammation upon allergen challenge, providing important in vivo functional data to support our GWAS finding. Targeting these pathways to promote the clearance of chronic inflammation in persistent disease may represent an exciting therapeutic prospect.

Appendix 1

Description of cohorts

The STELAR consortium (Custovic et al., 2015) brings together five UK population-based birth cohorts as described below. Informed consent was obtained from parents, and study subjects gave their assent/consent when applicable. Data were harmonised and imported into Asthma eLab web-based knowledge management platform to facilitate joint analyses (https://www.asthmaelab.org) (Custovic et al., 2015).

ALSPAC

ALSPAC is a birth cohort study established in 1991 in Avon, UK (Boyd et al., 2013; Fraser et al., 2013). Pregnant women with expected dates of delivery 1 April 1991 to 31 December 1992 were invited to take part in the study. The initial number of pregnancies enrolled is 14,541. Of these initial pregnancies, there was a total of 14,676 foetuses, resulting in 14,062 live births and 13,988 children who were alive at 1 year of age.

When the oldest children were approximately 7 years of age, an attempt was made to bolster the study with eligible cases who had failed to join originally. As a result, when considering variables collected from the age of 7 onwards (and potentially abstracted from obstetric notes) there are data available for more than the 14,541 pregnancies mentioned above. The number of new pregnancies not in the initial sample (known as Phase I enrolment) that are currently represented on the built files and reflecting enrolment status at the age of 24 is 913 (456, 262, and 195 recruited during Phases II, III, and IV, respectively), resulting in an additional 913 children being enrolled. The phases of enrolment are described in more detail in the cohort profile paper and its update. The total sample size for analyses using any data collected after the age of 7 is therefore 15,454 pregnancies, resulting in 15,589 foetuses. Of these 14,901 were alive at 1 year of age.

Ethical approval: Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committees. All self-completion questionnaire content is approved by the ALSPAC Ethics and Law Committee. Ethics protocols’ numbers: Initial approval Bristol and Weston Health Authority: E1808 Children of the Nineties: Avon Longitudinal Study of Pregnancy and Childhood (ALSPAC) (28 November 1989). Southmead Health Authority: 49/89 Children of the Nineties – ‘ALSPAC’ (5 April 1990). Frenchay Health Authority: 90/8 Children of the Nineties (28 June 1990).

Informed consent for the use of data collected via questionnaires and clinics was obtained from participants following the recommendations of the ALSPAC Ethics and Law Committee at the time. Data dictionary: The study website contains details of available data through a fully searchable data dictionary: http://www.bristol.ac.uk/alspac/researchers/our-data/.

We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses.

MAAS

MAAS is an unselected birth cohort study established in 1995 in Manchester, UK (Custovic et al., 2002). It consists of a mixed urban-rural population within 50 square miles of South Manchester and Cheshire, UK, located within the maternity catchment area of Wythenshawe and Stepping Hill Hospitals. All pregnant women were screened for eligibility at antenatal visits (8–10th week of pregnancy). Of the 1499 couples who met the inclusion criteria (≤10 weeks of pregnancy, maternal age ≥18 years, and questionnaire and skin prick data test available for both parents), 288 declined to take part in the study and 27 were lost to follow-up between recruitment and the birth of a child. A total of 1184 children were born into the study between February 1996 and April 1998. They were followed prospectively for 19 years to date and attended follow-up clinics for assessments, which included lung function measurements, skin prick testing, biological samples (serum, plasma, and urine), and questionnaire data collection. The study was approved by the North West – Greater Manchester East Research Ethics Committee. Ethics protocols’ numbers: ERP/94/032 Up to 5 years. Allergen avoidance, primary prevention, genetics, sRaw age 3 and 5; SOU/00/259 5 years; ERP/95/137 Exposure to pet allergens, atopy, genetics; ERP/97/023 IFWIN, genetics; 03/SM/400 8 years; 06/Q1403/142 10–12 years; 11/NW/0228 13–15 years; 14/NW/1309 18+ years.

SEATON

SEATON is an unselected birth cohort study established in 1997 in Aberdeen, UK, which was designed to explore the relationship between antenatal dietary exposures and asthma outcomes in childhood (Martindale et al., 2005). Two-thousand healthy pregnant women attending an antenatal clinic, at median 12 weeks gestation, were recruited. An interviewer administered a questionnaire to the women and atopic status was ascertained by skin prick test (SPT). The cohort included 1924 children born between April 1998 and December 1999. Participants were recruited prenatally and followed up by self-completion questionnaire to 15 years of age using postal questionnaires to record the presence of asthma and allergic diseases. Lung function measurements and SPT to common allergens were performed at 5, 10, and 15 years. The study was approved by the North of Scotland Research Ethics Committee. Ethics protocol REC reference: 13/NS/0108; protocol number: 2/048/13; amendment number: AM03.

Ashford

The Ashford study is an unselected birth cohort study established in 1991 in Ashford, UK (Atkinson et al., 1999). It included 642 children born between 1992 and 1993. Participants were recruited prenatally and followed to age 14 years. Detailed standardised questionnaires were administered at each follow-up to collect information on the natural history of asthma and other allergic diseases. Lung function measurements and SPT was carried out at 5, 8, and 14 years of age. In 2015, the study children aged 20 were sent a self-completion questionnaire, which was returned by 60% of the participants. The Asthma in Ashford study was reviewed by the Imperial College Research Ethics Committee on 11 November 2014. On 8 January 2015 the Joint Research Compliance Office granted full approval of the study on the basis described in the revised documents. ICREC reference: 14|C2288.

The IOW cohort

IOW is an unselected birth cohort study established in 1989 on the IOW, UK (Arshad et al., 2018; Kurukulaaratchy et al., 2002; Kurukulaaratchy et al., 2003). After the exclusion of adoptions, perinatal deaths, and refusal for follow-up, written informed consent was obtained from parents to enrol 1456 newborns born between 1 January 1989 and 28 February 1990. Follow-up-up assessments were conducted to 26 years of age to prospectively study the development of asthma and allergic diseases. At each follow-up, validated questionnaires were completed by the parents. Additionally, the SPT was performed on 980, 1036, and 853 participants at 4, 10, and 18 years of age to check allergic reactions to common allergens. At 10, 18, and 26 years, spirometry and methacholine challenge tests were performed to diagnose lung problems. Ethics protocols’ numbers: Ethics approval for the IoW cohort was originally given by the Isle of Wight Local Research Ethics Committee (now named the National Research Ethics Service, NRES Committee South Central – Southampton B) in 1989 and at each subsequent follow-up (1, 2, and 4 years) (this is pre ‘numbers’); age 10 follow-up (including DNA and genotyping): Isle of Wight Health Authority Local Research Ethics Committee 18/98; age 18 follow-up (including DNA and genotyping): Isle of Wight, Portsmouth & South East Hampshire Research Ethics Committee 06/Q1701/34.

Definition of variables

A list of all variables used in the current study, per cohort, is shown in Appendix 1—table 1.

Demographic, exposures and outcomes

Postal questionnaires were used in ALSPAC and SEATON, while interviewer-administered questionnaires were employed in other cohorts.

Parental history of asthma, eczema, and hay fever was defined based on the responses given to the question ‘have you (and/or your partner) ever had asthma/eczema/hay fever’. Maternal and paternal smoking were defined based on the response given to the question ‘do you (or does your partner) smoke’, administered during pregnancy. Low birth weight was defined as birth weight less than 2500 g based on NHS birth records.

Asthma in MAAS was defined as a case if positive for two of the following criteria: doctor diagnosis of asthma in the past 12 months, current wheeze in the last 12 months, doctor prescription for asthma. Asthma in ALSPAC was defined as a mothers’ report of doctor ever diagnosis of asthma.

Current wheeze in MAAS was defined as a questionnaire report to the question ‘have you wheezed in the last 12 months’ upon attendance at a follow-up clinic. Current wheeze in ALSPAC was defined as a mothers’ report to the question ‘has your child had any wheezing or whistling in the last 12 months?’.

Asthma medication in ALSPAC was defined as a mothers’ report to the question ‘has your child taken any asthma medication in the last 12 months?’. Lower respiratory hospital admissions: Data on hospital admissions in MAAS were obtained by manually inspecting the general practice (GP) records for each individual.

Early-life risk factors were divided into four groups according to timing of exposure: maternal and child characteristics (gender, maternal smoking during pregnancy, and maternal history of asthma), perinatal (low birth weight adjusted for gestational age), environmental (pet ownership, smoke exposure after birth), and allergic sensitisation (defined based on positive SPT to cat, HDM, or grass) variables.

Primary outcome: joint wheeze phenotypes

We used LCA to identify longitudinal trajectories of wheeze (Oksel et al., 2019a) based on pooled analysis among 15,941 children with at least two observations on wheezing at five time periods that were approximately shared across all cohorts: infancy (½–1 year); early childhood (2–3 years); pre-school/early school age (4–5 years); middle childhood (8–10 years); and adolescence (14–18 years). Cohort-specific definitions and other variables derived from the questionnaires are provided in Appendix 1—table 2.

To control for cohort-specific variation, cohort ID was included in the LCA model as an additional predictor by transforming the five-category variable into a set of four dummy variables and including them as covariates. The largest cohort, ALSPAC, was treated as the non-coded category to which all other cohorts were compared. The expectation maximisation algorithm was used to estimate relevant parameters, with 100,000 iterations and 500 replications.

To assess model fit, we used (1) the Bayesian information criterion (BIC), (2) the Akaike information criterion (AIC), (3) Lo-Mendell-Rubin likelihood ratio test, (4) bootstrapped likelihood ratio, and (4) quality of classification certainty (model entropy). The BIC is an index used in Bayesian statistics to choose among a set of competing models; the model with the lowest BIC is preferred. Using the lowest BIC as a selection criterion, the best fitting model was chosen as the five-class solution with a nominal covariate (BIC:31340). Analyses were carried out using Mplus 8, R (https://www.r-project.org/) and Stata 14 (StataCorp, College Station, TX, USA).

Based on the statistical fit, a five-class solution was selected as the optimal model (Oksel et al., 2019a), and the classes (wheeze phenotypes) were labeled as: (1) never/infrequent wheeze (52.4%); (2) early-onset pre-school remitting wheeze (18.6%), with high prevalence of wheeze during infancy, decreasing to 20% around early-childhood and to less than 10% afterwards; (3) early-onset middle-childhood remitting wheeze (9.8%), with early-onset wheeze and peak prevalence in early-childhood (~70%), and diminishing by middle-childhood (<5%); (4) early-onset persistent wheeze (10.4%) with 58% wheeze prevalence during infancy, and prevalence between 70–80% thereafter; (5) late-onset wheeze (8.8%) with very low prevalence until middle childhood, increasing rapidly to 55% in adolescence. These latent classes were used in the subsequent GWAS.

Minimising bias and missing data effects

Extracted from reference (Oksel et al., 2019a): “One of the advantages of our multicohort approach is that individual studies that might not provide conclusive evidence to make inference about the general population because of cohort specific effects and biases can contribute to revealing a more accurate picture when integrated together. The integration of five cohorts and their pooled analysis enhanced the credibility and generalizability of the phenotyping results to the U.K. population. A further advantage is to minimize the study-specific biases (including cohort specific effects, attrition effects, different recruitment strategies, and geographic factors) affecting the certainty of allocation of individuals to each latent class, while maximizing the benefits of individual cohort studies (e.g., potentially important risk factors and outcomes are captured in some, but not all cohorts).”

“Another strength of pooling cohort data is that a multicohort design allowed us to analyze a large sample with complete data on wheeze from birth to adolescence, thus increasing statistical power to detect less prevalent phenotypes.” However, “The optimal solution in the model using 15,941 children (allowing for missing data) remained five classes (see Table E3, Figure E1), and was very similar to that derived from a complete data set.” We used results from the larger sample, that is individuals with at least two observations of wheezing, to assign individuals to their most likely wheezing phenotype and used this as our primary outcome in this study.

Included vs. excluded participants

Related and non-European individuals were excluded as well as those individuals with missing genetic data.

In ALSPAC, 11,176 individuals had data on wheezing phenotypes, of these 6833 were white unrelated and had genetic data. We found more children from mothers who smoked during pregnancy in the excluded sample compared to the included sample; no difference in gender, maternal history of asthma, current wheezing at 8 or 15 years, and small evidence for more asthma ever and current medication at 8 years in the excluded sample (Appendix 1—table 4).

In MAAS, 1150 individuals had data on wheezing phenotypes, of these 887 were white unrelated and had genetic data. We found no difference in children from mothers who smoked during pregnancy in the excluded sample compared to the included sample; no difference in gender, maternal history of asthma or current wheezing at both 8 and 16 years. There was small evidence for more asthma ever and current medication at 8 years in the excluded sample (Appendix 1—table 4).

In SEATON, 1535 individuals had data on joint wheezing phenotypes, of these 548 were white unrelated and had genetic data. We found evidence for more children from mothers who smoked during pregnancy in the excluded sample compared to the included sample; and more males in the excluded sample. There was no difference in maternal history of asthma or current wheezing, asthma ever or current medication at both 10 and 15 years in the excluded sample compared to the included sample (Appendix 1—table 4).

In Ashford 620 individuals had data on joint wheezing phenotypes, of these 348 were white unrelated and had genetic data. We found evidence for more children from mothers who smoked during pregnancy in the excluded sample compared to the included sample; no difference in gender, maternal history of asthma, or asthma ever. There was small evidence for less current wheezing at 8 years, or current medication at 8 years in the excluded sample compared to the included sample (Appendix 1—table 4).

In IOW, 1460 individuals had data on joint wheezing phenotypes, of these 952 were white unrelated and had genetic data. We found evidence for more children from mothers who smoked during pregnancy in the excluded sample compared to the included sample; no difference in gender, maternal history of asthma, asthma ever at 10 and 18 years in the excluded sample compared to the included sample. There was small evidence for more children with current wheeze and medication at 8 years in the included sample compared to the included sample (Appendix 1—table 4).

Appendix 1—table 1
Definition of variables in each of the five Study Team for Early Life Asthma Research (STELAR) birth cohorts.
VariableDefinition
Cohort: ALSPAC
Mother – asthmaHave you ever had asthma? (recruitment)
Mother smokingMother smoked when expecting (recruitment)
Doctor-diagnosed asthma everHas a doctor ever said that your child has asthma? (years 8 and 14)
Current wheezingTwo questions combined: Occurrence of wheezing and/or wheezing with whistling on the chest in the last 12 months (year ½, 2½, 4¾, 8½, and 14)
Current asthma medicationAsthma medication in the last 12 months (years 8½ and 14)
Current rhinitisChild had problem with sneezing/runny nose without cold/flu in last 12 months (years 7 and 16½)
Current hay feverChild had hay fever in last 12 months (years 10½ and 14)
Cohort: MAAS
Mother – asthmaHas a doctor ever told you that you had asthma? (recruitment)
Mother smokingDo you smoke − mother (recruitment)
Doctor-diagnosed asthma everHas your doctor ever told you that your child has or had asthma? (years 8 and 16)
Asthma everHas your child ever suffered from asthma (years 8 and 16)
Current wheezingHas your child had wheezing or whistling in the chest in the last 6/12 months (years 1, 3, 5, 8, and 16)
Current asthma medicationAsthma medication in the last 12 months (years 8 and 16)
Current rhinitisHas your child ever had a problem with sneezing, or a runny nose, or a blocked nose when he /she did not have a cold or the flu? (years 8 and 16)
Current hay feverDoes your child have hay fever now? (years 8 and 16)
Cohort: SEATON
Mother – asthmaDo you suffer from asthma? (recruitment)
Mother smokingWhich of the following best describes your smoking status? (recruitment)
Doctor-diagnosed asthma everHas your child ever suffered from asthma? If yes, has this been confirmed by a doctor? (years 10 and 15)
Asthma everHas your child ever suffered from asthma? (year 10); Have you ever suffered from asthma? (year 15)
Current wheezingHas your child had wheezing in the chest in the last 12 months (years 1, 2, 5, 10, and 15)
Current asthma medicationHas your child been prescribed medicines/inhalers for asthma in the last 12 months? (year 10); Have you been prescribed medicines/inhalers for asthma in the last 12 months? (year 15)
Current hay feverHas your child suffered from hay fever last 12 months? (years 10 and 15)
Cohort: Ashford
Mother – asthmaDo you have or have you ever been told you have asthma? (recruitment)
Mother smokingDo you smoke cigarettes? (recruitment)
Doctor-diagnosed asthma everHas your doctor ever told you that your child has or had asthma? (years 8 and 14)
Asthma everIn the past 12 months has your daughter suffered from asthma? (year 8); Has she/he ever suffered from asthma? (year 14)
Current wheezingWhich one best describes your child’s wheeze in past 12 months? ‘Yes’ (B:1–6, C:7+), ‘No’ (A:0) (years 1, 2, 5, 8, and 14)
Current asthma medicationOver the last 12 months has your daughter taken any of the following treatments (preventer, reliever, nebuliser, steroids) for asthma? (years 8 and 14)
Current rhinitisIn the last 12 months has your child had a problem with sneezing or a runny or blocked nose? (years 8 and 14)
Current hay feverIn your opinion does your child have hay fever now? (year 8)
Has your child ever had hay fever? (year 14)
Cohort: IOW
Mother − asthmaDo you or have you suffered from asthma or wheezing (recruitment)
Mother smokingDo you smoke in the house? (recruitment)
Doctor-diagnosed asthma everAsthma cared for by hospital specialist/ GP or nurse (years 10, 18, and 26)
Asthma everChild ever had asthma (years 10 and 18)
Current wheezingPresence of wheeze since previous review (years 1, 2, 4, 10, and 18)
Asthma medication everChild ever had asthma treatment (year 18) combined with asthma treatment questions being asked at years 1, 2, 4, 10, and 18
Current rhinitisIn the past 12 months have you had a problem with sneezing, or a runny or blocked nose when you did not have a cold or the flu? (years 10, 18, and 26)
Appendix 1—table 2
The cohort-specific time points and sample size used to ascertain wheeze phenotypes.
Birth cohort:IOWMAASSEATONAshfordALSPAC
Year of birth19891995199719921991
QuestionnaireInterviewer-administeredInterviewer-administeredPostalInterviewer-administeredPostal
Data collection age (years)1, 2, 4, 10, 181, 3, 5, 8, 161, 2, 5, 10, 151, 2, 5, 8, 14½, 2½, 4¾, 8½, 14
No. of children with ≥2 observations on wheezing at five selected time points14601150153562011,176
Appendix 1—table 3
Characteristics of the participants in Study Team for Early Life Asthma Research (STELAR) cohorts included in this analysis (restricted to individuals with genetic data).

Numbers are N (%) except for age, where we report mean (SD).

ALSPACMAASSEATONAshfordIOW
N=6,833(71.4%)N=887(9.3%)N=548(5.7%)N=348(3.6%)N=952(9.9%)
Males3492 (51.1)475 (53.6)260 (47.5)179 (51.4)466 (49.0)
Maternal history of asthma748 (11.5)120 (13.5)77 (14.1)49 (14.1)106 (11.2)
Maternal smoking1423 (22.1)122 (13.8)107 (19.5)52 (14.9)217 (23.1)
Wheeze phenotypes
Never/infrequent4331 (63.4)506 (57.1)332 (60.6)145 (41.7)573 (60.2)
Early-onset persistent656 (9.6)133 (15.0)36 (6.6)41 (11.8)77 (8.1)
Early-onset pre-school remitting1076 (15.8)145 (16.4)117 (21.4)145 (41.7)0
Early-onset mid-childhood remitting474 (6.9)48 (5.4)13 (2.4)13 (3.7)55 (5.8)
Late-onset296 (4.3)55 (6.2)50 (9.1)4 (1.2)247 (26.0)
7–8 years14–15 years8 years16 years10 years15 years8 years14 years10 years18 years
Age mean (SD) in years8.7 (0.3)15.4 (0.3)7.98 (0.16)16.09 (0.62)10.15 (0.18)15.09 (0.28)7.97 (NA)13.95 (NA)9.98 (0.27)17.87 (0.59)
Doctor-diagnosed asthma ever*1060 (19.7)796 (23.2)198 (23.9)198 (30.0)86 (16.0)80 (19.5)75 (21.6)83 (23.9)350 (40.9)255 (28.6)
Asthma everNANA193 (22.8)192 (29.5)87 (16.2)66 (21.9)54 (15.6)65 (18.7)194 (20.9)264 (29.3)
Current wheeze683 (12.5)306 (9.0)150 (17.6)112 (16.9)67 (12.4)63 (15.5)54 (15.6)54 (15.5)190 (20.4)227 (25.1)
Current asthma medication695 (12.9)361 (10.6)141 (16.5)114 (17.1)68 (12.6)58 (14.0)50 (14.41)49 (14.1)41 (11.81)38 (10.9)
  1. *

    DDA ever not available in IOW, we used asthma cared for by hospital specialist/ GP or nurse as proxy.

Appendix 1—table 4
Comparison of included vs. excluded participants in the five cohorts at different ages.
ALSPACNIncludedNExcludedp-Value
Males (%)68333492 (51.1)43432269 (52.2)0.24
Maternal history asthma (%)6497748 (11.5)4038453 (11.2)0.64
Maternal smoking-pregnancy (%)64381423 (22.1)40191167 (29.0)<0.001
At 7.5–8.5 yearsAt 14–15 years
ALSPACNIncludedNExcludedp-ValueNIncludedNExcludedp-Value
Age mean (SD) years51398.7 (0.3)18728.7 (0.3)<0.001388515.4 (0.3)123715.5 (0.4)<0.001
Current wheeze (%)5453683 (12.5)2579344 (13.3)0.3083419306 (9.0)1078105 (9.7)0.432
Asthma ever (%)53771060 (19.7)2605562 (21.6)0.0533425796 (23.2)1079279 (25.9)0.079
Current asthma medication (%)5379695 (12.9)2529368 (14.6)0.0473400361 (10.6)1077134 (12.4)0.096
MAASNIncludedNExcludedp-Value
Males (%)887475 (53.6)263149 (56.7)0.38
Maternal history asthma (%)886120 (13.5)25945 (17.4)0.12
Maternal smoking* (%)884122 (13.8)26047 (18.1)0.09
At 8 yearsAt 16 years
MAASNIncludedNExcludedp-ValueNIncludedNExcludedp-Value
Age mean (SD) years8277.98 (0.16)1498.00 (0.21)0.3160516.09 (0.62)5915.98 (0.60)0.20
Current wheeze (%)853150 (17.6)17235 (20.4)0.39664112 (16.9)8215 (18.3)0.11
Asthma ever (%)845193 (22.8)17352 (30.1)0.043651192 (29.5)7928 (35.4)0.28
Current asthma medication (%)855141 (16.5)17343 (24.9)0.009666114 (17.1)8314 (16.9)0.96
SEATONNIncludedNExcludedp-Value
Males (%)548260 (47.5)987525 (53.2)0.031
Maternal history asthma (%)54877 (14.1)985161 (16.4)0.24
Maternal smoking* (%)548107 (19.5)987276 (28.0)<0.001
At 10 yearsAt 15 years
SEATONNIncludedNExcludedp-ValueNIncludedNExcludedp-Value
Age mean (SD) years54810.15 (0.18)98710.23 (0.16)<0.00154515.09 (0.28)91615.11 (0.26)0.20
Current wheeze (%)54167 (12.4)37642 (11.2)0.5840763 (15.5)31048 (15.5)0.99
Asthma ever (%)53787 (16.2)37453 (14.2)0.4040966 (21.9)30285 (20.8)0.73
Current asthma medication (%)54268 (12.6)37839 (10.3)0.3041458 (14.0)30934 (11.0)0.23
AshfordNIncludedNExcludedp-Value
Males (%)348179 (51.4)272153 (56.3)0.23
Maternal history asthma (%)34849 (14.1)27238 (14.0)0.97
Maternal smoking* (%)34852 (14.9)27061 (22.6)0.015
At 8 yearsAt 14 years
AshfordNIncludedNExcludedp-ValueNIncludedNExcludedp-Value
Age mean (SD) years348NA272NANA348NA272NANA
Current wheeze (%)34754 (15.6)24625 (10.2)0.0634854 (15.5)15018 (12.00)0.31
Asthma ever (%)34754 (15.6)24638 (15.5)0.9734865 (18.7)15025 (16.7)0.59
Current asthma medication (%)34750 (14.41)24622 (8.9)0.0534849 (14.1)15016 (10.7)0.30
IOWNIncludedNExcludedp-Value
Males (%)952466 (49.0)508275 (54.1)0.06
Maternal history asthma (%)946106 (11.2)50552 (10.3)0.60
Maternal smoking* (%)941217 (23.1)502147 (29.3)0.01
At 10 yearsAt 18 years
IOWNIncludedNExcludedp-ValueNIncludedNExcludedp-Value
Age mean (SD) years9329.98 (0.27)42610.04 (0.31)<0.00191417.87 (0.59)38918.14 (0.67)<0.001
Current wheeze (%)932190 (20.4)42669 (16.2)0.07903227 (25.1)37758 (15.4)<0.002
Asthma ever (%)930194 (20.9)42580 (18.8)0.39900264 (29.3)385108 (28.1)0.64
Current asthma medication (%)34741 (11.81)24615 (6.10)0.0234838 (10.9)15013 (8.7)0.45
Appendix 1—figure 1
Download asset Open asset
Flowchart of individuals included in the final meta-analysis.

Appendix 2

Genotyping and imputation

ALSPAC

Participants were genotyped using the Illumina HumanHap550 quad genome-wide SNP genotyping platform (Illumina Inc, San Diego, CA, USA) by the Wellcome Trust Sanger Institute (WTSI; Cambridge, UK) and the Laboratory Corporation of America (LCA, Burlington, NC, USA), using support from 23andMe. Haplotypes were estimated using ShapeIT (v2.r644) which uses relationship information to improve phasing accuracy. The phased haplotypes were then imputed to the Haplotype Reference Consortium (HRCr1.1, 2016) panel (Loh et al., 2016) of approximately 31,000 phased whole genomes. The HRC panel was phased using ShapeIt v2, and the imputation was performed using the Michigan imputation server.

MAAS

In MAAS, we used the Illumina 610 quad genome-wide SNP genotyping platform (Illumina Inc, San Diego, CA, USA). Prior to imputation samples were excluded on the basis of gender mismatches; minimal or excessive heterozygosity, genotyping call rates of <97%. SNPS were excluded if they had call rates of <95%, minor allele frequencies of <0.5%, and HWE p<3 × 10–8. Prior to imputation each chromosome was pre-phased using EAGLE2 (v2.0.5) (Loh et al., 2016) as recommended by the Sanger imputation server (McCarthy et al., 2016). We then imputed with PBWT (Durbin, 2014) with the Haplotype Reference Consortium (release 1.1) of 32,470 reference genomes (McCarthy et al., 2016) using the Sanger Imputation Server.

IOW, SEATON, and Ashford

IOW, SEATON, and Ashford were genotyped using the illumina Infinium Omni2.5–8 v1.3 BeadChip genotyping platform (Illumina Inc, San Diego, CA, USA). Genotype QC and imputation was carried out as described for MAAS.

Exclusions

Individuals were excluded on the basis of gender mismatches; minimal or excessive heterozygosity; disproportionate levels of individual missingness (>3%), insufficient sample replication (IBD <0.8), or evidence of cryptic relatedness (IBD >0.1). Following imputation, SNPs with a minor allele frequency of <1%, a call rate of <95%, evidence for violations of Hardy-Weinberg equilibrium (p<5E-7), or imputation quality measure (MaCH-Rsq or IMPUTE-info score)<0.40 were excluded. All individuals with non-European ancestry and siblings were removed.

GWAS meta-analysis

GWASs of the joint wheezing phenotypes were performed independently in ALSPAC, MAAS, and the combined IOW-SEATON-Ashford (combined as they were genotyped on the same platform, at the same time, and quality-controlled and imputed together). All genetic data were imputed to a new Haplotype Reference Consortium panel. This comprises around 31,000 sequenced individuals (mostly European), so the coverage of European haplotypes is much greater than in other panels. As a consequence, we expect to improve imputation accuracy, particularly at lower frequencies.

We used SNPTEST v2.5.2 (Marchini and Howie, 2010) with a frequentist additive multinomial regression model (-method newml, never/infrequent wheeze as the reference) to investigate the association between SNPs and wheezing phenotypes. No covariates were included in the model and only individuals of European descent were included in this analysis. A meta-analysis of the three GWASs, including 5887 controls and 943 cases for early-onset persistent, 1482 cases for early-onset remitting, 603 cases for mid-childhood onset remitting, and 652 cases for late-onset wheeze, was performed using METAL (Willer et al., 2010) with a total of 8,057,852 SNPs present. We used the option SCHEME STDERR in METAL to implement an effect size-based method weighted by each study-specific standard error in a fixed-effects model. We performed clumping to keep only one representative SNP per LD block and used LZPs to short-list independent SNPs for further annotation.

LD clumping, pre-selection, and gene annotation

LD clumping was performed for all SNPs with p-value <10–5 for at least one wheezing phenotype. In order to avoid redundancy between SNPs and to ensure associations are independent, we used significance thresholds of 0.05 for index and clumped SNPs (--clump-p1 0.05, --clump-p2 0.05), LD threshold of 0.80 (--clump-r2 0.80) and physical distance threshold of 250 kb (--clump-kb 250). European 1000 Genome data were used to infer LD structure.

LZPs (http://locuszoom.org/) (Pruim et al., 2010) were used for close inspection of all independent signals. Loci showing a peak with different colour dots (possibly indicating more than one causal variant) were short-listed for further annotation. SNPnexus database (https://www.snp-nexus.org/v4/) (Dayem Ullah et al., 2018) was used to annotate the overlapping, upstream and downstream genes; the GWAS Catalog (by SNP and then gene) (https://www.ebi.ac.uk/gwas/search), GeneCards (https://www.genecards.org/) (Stelzer et al., 2016), database, and phenoscanner (http://www.phenoscanner.medschl.cam.ac.uk/) were used to further explore previously associated relevant phenotypes and gene function. Lead SNPs were looked in https://www.regulomedb.org/ to assess potential functionality.

Genetic control

The genomic inflation factor (λ) was calculated using the scipy.stats.chi2 module in Python. The chi-squared test statistics from the meta-analysis p-values were first obtained. Then, the observed median chi-squared statistic from the calculated chi-squared test statistics were calculated. Finally, the genomic inflation factor (λ) was derived by dividing the observed median chi-squared statistic by the expected median chi-squared statistic.

Heterogeneity scatter plots

Heterogeneity scatter plots were based on filtering signals for each pair of wheezing phenotypes. For example, for group1=persistent and group2=early-onset mid-childhood remitting wheezing.

If group1 has a p-value <10–5 and group2 has a p-value >0.05, and group1 has a negative effect size (beta) while the lower bound of group2’s effect size (beta - CI) is greater than group1’s effect size, then we classified the result as group1 specific. If group1 has a p-value <10–5 and group2 has a p-value >0.05, and group1 has a positive effect size (beta) while the upper bound of group2’s effect size (beta+CI) is less than group1’s effect size, then we classified the result as group1 specific.

If group2 has a p-value <10–5 and group1 has a p-value >0.05, and group2 has a negative effect size (beta) while the lower bound of group1’s effect size (beta - CI) is greater than group2’s effect size, then we classified the result as group2 specific. If group2 has a p-value <10–5 and group1 has a p-value >0.05, and group2 has a positive effect size (beta) while the upper bound of group1’s effect size (beta+CI) is less than group2’s effect size, then we classified the result as group2 specific.

If both group1 and group2 have p-values <10–5, and their effect sizes (betas) have the same sign (i.e., both positive or both negative), then we classified the result as ‘Same direction’.

If both group1 and group2 have p-values <10–5, and their effect sizes (betas) have opposite signs (i.e., one positive and one negative), then we classified the result as ‘Opposite direction’.

Gene expression in whole blood and lung tissues

The top independent SNPs associated with each of the wheeze phenotypes were assessed for their association with cis- and trans-acting gene expression (mRNA) in whole blood and lung tissues. We identified potential eQTL signals using Genotype-Tissue Expression database (https://gtexportal.org) using the European reference panel.

Appendix 2—table 1
List of 134 independent single nucleotide polymorphisms (SNPs) identified after clumping and associated with at least one wheezing phenotype (p<10–5).
CHRSNPBPShort-listed after inspection of locus zoom plot
Persistent wheezing
1rs4620530240063821Yes
2rs133984887142199Yes
2rs7706232353049017No
2rs6543291106011626Yes
3rs77655717128737320Yes
4rs778226211008212Yes
4rs76806081050437Yes
4rs115228498142969757Yes
4rs145937716143192224No
5rs1164941157736317Yes
5rs7870148395680422No
6rs1380999417654240No
6rs934640471606613No
6rs143979498151040328No
7rs76871421105676144Yes
8rs59670576128555771No
9rs11693312027458652No
9rs7526065475788108Yes
9rs11684966475820902Yes
9rs143481506139515723No
10rs7088157100038964Yes
11rs1124745745885773Yes
11rs116861530116962661Yes
13rs798235073106322No
13rs17461573106711373No
14rs110568356213787Yes
15rs220271449811991Yes
15rs11754021484338642Yes
17rs1767619137949924Yes
17rs7902687237965932Yes
17rs479540038067020Yes
17rs103146038072247Yes
17rs5619942138090808Yes
17rs479540638100134Yes
17rs7283297238110575Yes
17rs479482138124203Yes
17rs5984358438124892Yes
18rs11181299330353181No
19rs48043118615589Yes
19rs20136948616392Yes
19rs735015458620823Yes
19rs1116449458625081Yes
22rs599417017615213Yes
22rs3490237017632194Yes
Early-onset remitting wheezing
1rs12730098212427488Yes
1rs75639566233019116No
2rs288006617107219Yes
2rs1018026817126699Yes
3rs11503179686691640No
3rs3861377173317378Yes
3rs10513743176022304Yes
5rs1007525375548246Yes
5rs1252088484406634No
6rs11747729792565052No
6rs2453395166286532Yes
7rs5602786950072919No
7rs473056178531705Yes
7rs7314497678586112Yes
7rs6725932178686582Yes
7rs146771277154438861No
9rs1075825934392908Yes
11rs712899471242209No
11rs72994149106837223Yes
12rs11736725693508478No
13rs287294857442480Yes
13rs7352765457447994Yes
13rs215150482291577No
15rs11696688647043587Yes
15rs11756552747342882Yes
Mid-childhood onset remitting wheezing
1rs35725789159207367Yes
1rs146141555159227423Yes
1rs146575092159374228Yes
1rs140877050220848829No
1rs72745905223451086No
2rs759555336127878Yes
2rs14500750350688324No
2rs654606864583398No
2rs17387431206651315No
2rs144791928236963432No
3rs343159998969653Yes
3rs11524577099209128No
3rs146961758194285978Yes
4rs138794367103859545Yes
5rs11571940277538102Yes
6rs7602639947531792No
7rs73172838154842348No
8rs112631708134500083No
9rs7275235698094970No
13rs11319538446333770No
13rs960221884139813Yes
13rs6196036684144202Yes
13rs7458992784208697Yes
13rs221072684492936Yes
13rs439047684598570Yes
14rs11744346473460284Yes
16rs7282081481916262No
17rs19052669712274299No
18rs7528653426206826No
18rs13888808663591085No
18rs7655153571879807No
19rs7749644419192132No
20rs60775149302948Yes
Late-onset wheezing
1rs943966918859049Yes
1rs205103957067560Yes
1rs7267364280727443Yes
2rs14755711719778063No
2rs140983998111402871Yes
2rs117617447123387601No
2rs13025116127505482No
2rs148008098128633620Yes
3rs407272924780393Yes
3rs14396066631227943No
3rs467710272193991No
3rs145629570113422516Yes
3rs113643470141728174Yes
4rs1747201548467594Yes
7rs117660982149438923Yes
7rs118027705150456728Yes
7rs139489493150481499Yes
7rs144271668157934780Yes
8rs99018289976447Yes
9rs7911096214432953Yes
10rs932546082492323No
10rs789610691196402Yes
10rs115465993109372900No
11rs1693564341395746No
11rs141958628119083284Yes
14rs11336366069410278No
15rs13913426544923960Yes
15rs14386203084922146Yes
16rs1133903671118849Yes
16rs478802528003221Yes
18rs7291826451009510No
22rs13349848913809Yes
Appendix 2—figure 1
Download asset Open asset
Zoom locus plots for short-listed independent top hits for persistent wheezing.
Appendix 2—figure 2
Download asset Open asset
Zoom locus plots for short-listed independent top hits for early-onset pre-school remitting wheezing.
Appendix 2—figure 3
Download asset Open asset
Zoom locus plots for short-listed independent top hits for early-onset mid-childhood remitting wheezing.
Appendix 2—figure 4
Download asset Open asset
Zoom locus plots for short-listed independent top hits for late-onset wheezing.

Appendix 3

Post-GWAS: rs75260654 (ANXA1)

Annotation and distribution

Information including chromosome, strand, clinical significance was retrieved from ENSEMBL using the R package biomaRt (Durinck et al., 2005; Durinck et al., 2009). The effects of rs75260654 on genomic features were predicted by querying the Ensembl VEP (McLaren et al., 2016) web tool.

rs75260654 distribution in the GRCh38.p13 build of the human genome across African, Asian, and European populations of the 1000 Genomes Project Phase 3 were accessed by querying the Ensembl (https://www.ensembl.org) web browser on 24 May 2021.

Promoter Capture

The Hi-C libraries were prepared from CD4+ T cells isolated from seven healthy individuals (two libraries per individual) from the MAAS cohort using the Arima-HiC kit (Arima Genomics). PCHi-C libraries were generated by capturing the restriction fragments (RF) overlapping the TSS of 18,775 protein-coding genes using the Agilent SureSelectXT HS Target Enrichment System according to the manufacturer’s protocols. The final design included 305,419 probes covering 13.476 Mb and 18,630 protein-coding genes. The RF overlapping the TSS (±1 RF, 3 RF per promoter) were captured with custom-designed biotinylated RNA baits. Libraries were sequenced to ~300 M 2×150 bp reads each (~600 M reads/individual). The 3’-end of the reads was quality trimmed with Sickle. The sequencing data were processed with the HiCUP pipeline to map the sequencing reads and eliminate experimental artefacts and PCR duplicates (Wingett et al., 2015). The BAM files from technical replicates were merged. Promoter interactions were called using the CHiCAGO pipeline (Cairns et al., 2016), which calls statistically significant interactions in PCHi-C data while accounting for noise and PCHi-C-specific bias. A CHiCAGO score >5 (soft-thresholded -log weighted p-value) was considered significant. To gain information from all the available data, the BAM files from all seven individuals were supplied as biological replicates in the analysis with CHiCAGO. Moreover, to increase power, RF were binned as follows: 10 consecutive RF that were not covered by the baits were binned together; the 3 baited fragments for each promoter were binned with 1 RF upstream and 1 downstream, totalling 5 fragments per promoter. If the bins for two consecutive promoters overlapped, these were binned together into a single larger bin. Publicly available ENCODE ATAC-seq (A549 cell line) and ChIP-seq (A549 cell line ENCFF900GVO) and POLR2A ChIP-seq (A549 cell line, ENCFF737ZKN) data and 18-state ChromHMM from the EpiMap Project (BSS00007) (Adsera et al., 2019) for A549 cell line were downloaded. The PCHi-C interactions of interest and their overlap with ATAC-seq and ChIP-seq peaks, and putative enhancers from the 18-state ChromHMM model were visualised using the Sushi R package.

eQTL catalogue lookup

We queried the eQTL catalogue (https://www.ebi.ac.uk/eqtl/; accessed 6 May 2021) using tabix-0.2.6 to assess if rs75260654, rs116849664, or rs78320984 are eQTLs in studies that utilised the following cell types: lung, T cells, blood, monocytes, neutrophils, NK cells, fibroblasts, B cells, CD4+ T cells, CD8+ T cells, Th17 cells, Th1 cells, Th2 cells, Treg naive, Treg memory, CD16+ monocytes, cultured fibroblasts, EBV-transformed lymphocytes. We defined nominal significance as p ≤ 0.05.

Variant effect

Variant effect on tissue-specific gene expression, which is based on GTEx eQTL, was retrieved on May 24 from eQTL Ensembl database (https://www.ensembl.org/) and eQTLGene Consortium (https://www.eqtlgen.org/cis-eqtls.html). Using downloaded correlation of variant on tissue-specific gene expression from Ensembl, the relative effect of T allele on the ANXA1 expression across 86 tissue types was presented in scatter plot using R version 3.6.1 (Ihaka and Gentleman, 1996). To get information on the functional role of ANXA1, the top 30 interacting proteins and enrichment were retrieved from STRING database (Szklarczyk et al., 2019) into cystoscape for visualisation (Shannon et al., 2003).

Appendix 4

Results in context of literature

Previously relevant associated traits for each region/gene now associated with different wheezing phenotypes are presented in Appendix 5—tables 1–4.

Persistent wheeze

We identified two GWAS-significant loci: 17q21, p<5.5 × 10–9, and a novel locus on 9q21.13 (ANXA1), p<6.7 × 10–9. The remaining 17 loci (4.0×10–7 ≤ p-values ≤9.8 × 10–6) included regions previously associated with childhood asthma (1q43, 4p16.3, 4q31.21, 5p15.31, 7q22.3, 17q12), asthma and rhinitis (2p25.1), eosinophil count (3q21.3, 10q24.2, 11q23.3, 22q11.1), bronchial hyper-responsiveness (2q12.2), lung function (1q43, 3q21.3, 5q13.3, 15q25.2, 19p13.2), triglycerides measurement and/or glucose metabolism (11q23.3, 19p13.2, and 22q11.1), severe asthma (14q22.1), and severe asthma and insulin resistance (11p15.4). See Appendix 5—table 1.

Early-onset pre-school remitting wheeze

Among the regions associated with early-onset pre-school remitting wheeze, we identified loci previously associated with smoking (3q26.31, 7q21.11, and 15q21.1), waist circumference and obesity (1q32.3), asthma and/or BMI (5q13.3, 6q27, 7q21.11), allergic disease and atopy (7q21.11), and airway repair (2p24.2 and 9p13.3). See Appendix 5—table 2.

Early-onset mid-childhood remitting wheeze

Loci associated with this phenotype were previously associated with neutrophil counts (1q23.2, 3q29, 20p12.3-p12.2), eosinophil counts and allergic rhinitis (4q24), pollution and DNA methylation (2p22.3), atopy (3p25.3), food allergy (13q31.1), and BMI (3q29, 5q14.1). See Appendix 5—table 3.

Late-onset wheeze

Regions associated with late-onset wheeze were previously associated with adult-onset non-allergic asthma (3q13.2), asthma/allergic disease and allergy/atopic sensitisation (3q23, 7q36.1), asthma and/or allergy in adolescence (9p22.3, 16p12.1), late-onset asthma and obesity (2q13), lung function or body height (2q14.3, 3p24.2, 3q13.2, 15q25.2), lymphocyte count and asthma susceptibility (1p32.2), obesity and/or metabolic syndrome/dysfunction (1p36.13 and 22q13.32), eczema (7q36.3), insulin resistance (10q23.31), type 1 diabetes (8q21.3), alcohol drinking (15q15.3-q21.1), and sex hormone-binding globulin levels (11q23.3). See Appendix 5—table 4.

Appendix 4—table 1
References to previous relevant associated traits for each region/gene identified in early-onset persistent wheezing.
Early-onset persistent wheezing
Gene(s)LocusPrevious associated traitReference or source
CHRM31q43FEV1, FEV1/FVC, asthma – high priority drug targetPatel, K.R. et al. Targeting acetylcholine receptor M3 prevents the progression of airway hyperreactivity in a mouse model of childhood asthma. FASEB J 31, 4335–4346 (2017).
RNF144A2p25.1Asthma, allergy, childhood onset asthma, allergic rhinitisSchoettler, N. et al. Advances in asthma and allergic disease genetics: Is bigger always better? J Allergy Clin Immunol 144, 1495–1506 (2019).
FHL22q12.2Bronchial hyper-responsiveness, airway inflammation, novel gene associated with asthma severity in humanKurakula, K. et al. Deficiency of FHL2 attenuates airway inflammation in mice and genetic variation associates with human bronchial hyper-responsiveness. Allergy 70, 1531–44 (2015).
RAB7A3q21.3Eosinophil countGeneCards
RAB433q21.3Response to bronchodilator, FEV1/FEC ratioGWAS Catalog
RNF212, IDUA, DGKQ, SLC26A14p16.3AsthmaGautam, Y. et al. Comprehensive functional annotation of susceptibility variants associated with asthma. Hum Genet 139, 1037–1053 (2020).
INPP4B4q31.21Atopic asthmaSharma, M. et al. A genetic variation in inositol polyphosphate 4 phosphatase a enhances susceptibility to asthma. Am J Respir Crit Care Med 177, 712–9 (2008).
ADCY25p15.31Asthma×air pollution, childhood asthmaGref, A. et al. Genome-Wide Interaction Analysis of Air Pollution Exposure and Childhood Asthma with Functional Follow-up. Am J Respir Crit Care Med 195, 1373–1383 (2017).
CDHR37q22.3Childhood asthmaEverman, J.L. et al. Functional genomics of CDHR3 confirms its role in HRV-C infection and childhood asthma exacerbations. J Allergy Clin Immunol 144, 962–971 (2019).
ANXA19q21.13FEV1/FVC, response to bronchodilators in smokersLutz, S.M. et al. A genome-wide association study identifies risk loci for spirometric measures among smokers of European and African ancestry. BMC Genet 16, 138 (2015).
ANXA19q21.13Anti-inflammatory properties, strongly expressed in bronchial mast cellsVieira Braga FA et al. A cellular census of human lungs identifies novel cell states in health and in asthma. (2019).
ANXA19q21.13Potentially involved in epithelial airway repairLeoni, G. et al. Annexin A1-containing extracellular vesicles and polymeric nanoparticles promote epithelial wound repair. J Clin Invest 125, 1215–27 (2015).
R3HCC1L10q24.2Atopic eczema, psoriasisGWAS Catalog
R3HCC1L10q24.2Eosinophil count, BMIGeneCards
TRIM5, TRIM6, TRIM2211p15.4Severe asthma and insulin resistanceKimura, T. et al. Precision autophagy directed by receptor regulators - emerging examples within the TRIM family. J Cell Sci 129, 881–91 (2016).
SIK311q23.3Triglycerides, glucose metabolism, eosinophil countSun, Z. et al. The potent roles of salt-inducible kinases (SIKs) in metabolic homeostasis and tumorigenesis. Sig Transduct Target Ther 5 (2020).
KTN114q22.1Severe asthmaBigler, J. et al. A Severe Asthma Disease Signature from Gene Expression Profiling of Peripheral Blood from U-BIOPRED Cohorts. Am J Respir Crit Care Med 195, 1311–1320 (2017).
FAM227B5q13.3rs35251997 and FEV1, FEV1/FVCShrine, N. et al. New genetic signals for lung function highlight pathways and chronic obstructive pulmonary disease associations across multiple ancestries. Nat Genet 51, 481–493 (2019).
ADAMTSL315q25.2FEV1/FVCSakornsakolpat, P. et al. Genetic landscape of chronic obstructive pulmonary disease identifies heterogeneous cell-type and phenotype associations. Nat Genet 51, 494–505 (2019).
IKZF3, GSDMB, LRRC3C, GSDMA17q12Early-onset asthma, persistent wheezing (chr17q12-q21)Granell R et al. Examination of the relationship between variation at 17q21 and childhood wheeze phenotypes. J Allergy Clin Immunol. 2013 Mar;131(3):685–94.
MARCH2, HNRNPM, MYO1F19p13.2Triglycerides, HDL cholesterol, metabolic syndromeSajuthi, S.P. et al. Genetic regulation of adipose tissue transcript expression is involved in modulating serum triglyceride and HDL-cholesterol. Gene 632, 50–58 (2017).
MYO1F19p13.2FEV1 and FVCGeneCards
CECR522q11.1Triglycerides, eosinophil count, and body heightLiu, D.J. et al. Exome-wide association study of plasma lipids in >300,000 individuals. Nat Genet 49, 1758–1766 (2017).
Appendix 4—table 2
References to previous relevant associated traits for each region/gene identified in early-onset pre-school remitting wheezing.
Early-onset pre-school remitting wheezing
Gene(s)LocusPrevious associated traitReference or source
PPP2R5A1q32.3Waist circumference and obesityKim, H.J. et al. Combined linkage and association analyses identify a novel locus for obesity near PROX1 in Asians. Obesity (Silver Spring) 21, 2405–12 (2013).
FAM49A or CYRIA2p24.2Airway repair in non-atopic asthmaHoang, T.T. et al. Epigenome-wide association study of DNA methylation and adult asthma in the Agricultural Lung Health Study. Eur Respir J 56 (2020).
NLGN13q26.31SmokingDrgon, T. et al. Genome-wide association for nicotine dependence and smoking cessation success in NIH research volunteers. Mol Med 15, 21–7 (2009).
NAALADL23q26.31Suggestive association with severe asthma exacerbationsHerrera-Luis E et al. Genome-wide association study reveals a novel locus for asthma with severe exacerbations in diverse populations. Pediatr Allergy Immunol. 2021;32(1):106–115.
SV2C5q13.3BMI, diastolic blood pressureGeneCards
PDE10A6q27Birth weight, asthma, and BMIMelen, E. et al. Analyses of shared genetic factors between asthma and obesity in children. J Allergy Clin Immunol 126, 631–7 e1-8 (2010).
MAGI27q21.11Allergic diseases and atopyFreidin, M.B. et al. [Genome-wide association study of allergic diseases in Russians of Western Siberia]. Mol Biol (Mosk) 45, 464–72 (2011).
MAGI27q21.11SmokingQuach, B.C. et al. Expanding the genetic architecture of nicotine dependence and its shared genetics with multiple traits. Nat Commun 11, 5562 (2020).
MAGI27q21.11BMIGeneCards
MAGI27q21.11Airway wall thicknessGWAS Catalog
C9orf249p13.3Airway repairYoshisue, H. et al. Characterisation of ciliated bronchial epithelium 1, a ciliated cell-associated gene induced during mucociliary differentiation. Am J Respir Cell Mol Biol 31, 491–500 (2004).
GUCY1A211q22.3Systolic/diastolic blood pressureGeneCards
PRR20A/B/C/D/E13q21.1Systolic blood pressureGeneCards
SEMA6D15q21.1SmokingMinica, C.C. et al. Pathways to smoking behaviours: biological insights from the Tobacco and Genetics Consortium meta-analysis. Mol Psychiatry 22, 82–88 (2017).
Appendix 4—table 3
References to previous relevant associated traits for each region/gene identified in early-onset mid-childhood remitting wheezing.
Early-onset mid-childhood remitting wheezing
Gene(s)LocusPrevious associated traitReference
CADM3, FCER1A, MPTX1, OR10J11q23.2Neutrophil count, CRPBarreto, M. et al. Duffy phenotyping and FY*B-67T/C genotyping as screening test for benign constitutional neutropenia. Hematol Transfus Cell Ther (2020).
MRPL50P12p22.3PM 2.5 exposure level and global DNA methylation levelLiu, J. et al. Genetic variants, PM2.5 exposure level and global DNA methylation level: A multi-center population-based study in Chinese. Toxicol Lett 269, 77–82 (2017).
RAD183p25.3Atopy/SPTBouzigon, E. et al. Meta-analysis of 20 genome-wide linkage studies evidenced new regions linked to asthma and atopy. Eur J Hum Genet 18, 700–6 (2010).
MRPL50P13q29BMIKettunen, J. et al. Multicenter dizygotic twin cohort study confirms two linkage susceptibility loci for body mass index at 3q29 and 7q36 and identifies three further potential novel loci. Int J Obes (Lond) 33, 1235–42 (2009).
LSG13q29BMI, eosinophil, and neutrophil countGeneCards
TMEM44-AS1, TMEM44, ATP13A33q29Diastolic blood pressureGeneCards
SLC9B14q24Eosinophil countAschard, H. et al. Sex-specific effect of IL9 polymorphisms on lung function and polysensitization. Genes Immun 10, 559–65 (2009).
SLC9B14q24Allergic rhinitisHaagerup, A. et al. Allergic rhinitis--a total genome-scan for susceptibility genes suggests a locus on chromosome 4q24-q27. Eur J Hum Genet 9, 945–52 (2001).
AP3B15q14.1Vital capacity, BMIGeneCards, GWAS Catalog
RNU6-67P, SLITRK1, VENTXP2, UBE2D3P4, MTND4P113q31.1RNU6-67P/ rs976078: food allergyLiu, X. et al. Genome-wide association study of maternal genetic effects and parent-of-origin effects on food allergy. Medicine (Baltimore) 97, e0043 (2018).
ZFYVE114q24.2LDL cholesterol and systolic blood pressureGWAS Catalog
PLCB420p12.3-p12.2Neutrophil countOkada, Y. et al. Common variations in PSMD3-CSF3 and PLCB4 are associated with neutrophil count. Hum Mol Genet 19, 2079–85 (2010).
Appendix 4—table 4
References to previous relevant associated traits for each region/gene identified in late-onset wheezing.
Late-onset wheezing
Gene(s)LocusPrevious associated traitReference
KLHDC7A1p36.131p36.13: metabolic syndromeHoffmann, K. et al. A German genome-wide linkage scan for type 2 diabetes supports the existence of a metabolic syndrome locus on chromosome 1p36.13 and a type 2 diabetes locus on chromosome 16p12.2. Diabetologia 50, 1418–22 (2007).
PPAP2B, PRKAA21p32.2PRKAA2: lymphocyte count and asthma susceptibilityCusanovich, D.A. et al. The combination of a genome-wide association study of lymphocyte count and analysis of gene expression data reveals novel asthma candidate genes. Hum Mol Genet 21, 2111–23 (2012).
HMGB1P181p31.1Smoking, BMIGeneCards
ACOXL, BUB12q13ACOXL: later onset asthma and obesityZhu, Z. et al. Shared genetic and experimental links between obesity-related traits and asthma subtypes in UK Biobank. J Allergy Clin Immunol 145, 537–549 (2020).
AMMECR1L2q14.3Body height, blood protein, growth, bone, and heart alterationsMoyses-Oliveira, M. et al. Inactivation of AMMECR1 is associated with growth, bone, and heart alterations. Hum Mutat 39, 281–291 (2018).
RARB3p24.2FEV1/FVC, adult lung functionCollins, S.A. et al. HHIP, HDAC4, NCR3 and RARB polymorphisms affect fetal, childhood and adult lung function. Eur Respir J 41, 756–7 (2013).
KIAA2018, NAA50, SIDT1, CD2003q13.2SIDT1: FEV1/FVC, CD200: adult-onset non-allergic asthmaSiroux, V. et al. Genetic heterogeneity of asthma phenotypes identified by a clustering approach. Eur Respir J 43, 439–52 (2014).
TFDP2, XRN13q23XRN1: eosinophil count, 3q23: allergic disease and atopic sensitisationFreidin, M.B. et al. [Genome-wide association study of allergic diseases in Russians of Western Siberia]. Mol Biol (Mosk) 45, 464–72 (2011).
SLAIN2, SLC10A4, FRYL4p11FRYL: body height, age at menopauseGeneCards
KRBA1, ZNF4677q36.1Systolic blood pressureGWAS Catalog
GIMAP family, AOC1, LOC1053755667q36.1AOC1: CV disease, smoking, GIMAP family: autoimmune diabetes, asthma and allergyHeinonen, M.T. et al. GIMAP GTPase family genes: potential modifiers in autoimmune diabetes, asthma, and allergy. J Immunol 194, 5885–94 (2015).
PTPRN27q36.3EczemaBogari, N.M. et al. Whole exome sequencing detects novel variants in Saudi children diagnosed with eczema. J Infect Public Health 13, 27–33 (2020).
LOC1053756318q21.38q21.3: type 1 diabetesMukhopadhyay, N., Noble, J.A., Govil, M., Marazita, M.L. & Greenberg, D.A. Identifying genetic risk loci for diabetic complications and showing evidence for heterogeneity of type 1 diabetes based on complications risk. PLoS One 13, e0192696 (2018).
NFIB, ZDHHC219p22.39p22.3: asthma (mean age <16 years)Denham, S. et al. Meta-analysis of genome-wide linkage studies of asthma and related traits. Respir Res 9, 38 (2008).
SLC16A12, IFIT family, PANK110q23.31SLC16A12: body height, PANK1: insulin resistanceYang, L. et al. P53/PANK1/miR-107 signalling pathway spans the gap between metabolic reprogramming and insulin resistance induced by high-fat diet. J Cell Mol Med 24, 3611–3624 (2020).
CBL, CCDC84, MCAM11q23.3CBL: sex hormone-binding globulin levels; MCAM: blood protein levelsGWAS Catalog
SPG11, CTDSPL215q15.3-q21.1CTDSPL2: alcohol drinkingGWAS Catalog
ADAMTSL3, GOLGA6L4, UBE2Q2P815q25.2ADAMTSL3: FEV1/FVC, lean massKarasik, D. et al. Disentangling the genetics of lean mass. Am J Clin Nutr 109, 276–287 (2019).
SSTR5-AS1, CACNA1H16p13.3CACNA1H: eosinophil countGWAS Catalog
GSG1L16p12.116p12.1: current asthma and rhino-conjunctivitis at 10–15 yearsSottile, G. et al. An association analysis to identify genetic variants linked to asthma and rhino-conjunctivitis in a cohort of Sicilian children. Ital J Pediatr 45, 16 (2019).
FAM19A5 or TAFA522q13.32Obesity and metabolic dysfunctionRecinella L. et al. Adipokines: New Potential Therapeutic Target for Obesity and Metabolic, Rheumatic, and Cardiovascular Diseases. Front Physiol. 2020 Oct 30;11:578966

Appendix 5

Replication of ANXA1 top hits in PIAMA cohort

PIAMA cohort description

PIAMA (Prevention and Incidence of Asthma and Mite Allergy) is an ongoing birth cohort study. Details of the study design have been published previously (Brunekreef et al., 2002; Wijga et al., 2014). In brief, pregnant women were recruited from the general population through antenatal clinics in the north, west, and centre of the Netherlands in 1996–1997. The baseline study population consisted of 3963 newborns. Questionnaires were completed by the parents during pregnancy when the child was 3 months old, and then annually from 1 up to 8 years; at ages 11, 14, and 17 years, questionnaires were completed by the parents as well as the participants themselves.

LCA wheezing phenotypes

A six-class LCA model was identified including 3832 individuals with at least two observations of wheeze between 1 and 11–12 years of age. The identified classes were labelled: never/infrequent (2909, 75.91%), pre-school onset remitting (571, 14.90%), mid-childhood school remitting (108, 2.82%), intermediate onset remitting (106, 2.77%), school-age onset persisting (74, 1.93%), and continuous wheeze (64, 1.67%).

Replication analyses

We analysed associations between SNPs downstream of ANXA1 (Appendix 5—table 1, Appendix 5—figure 1) and continuous wheezing in PIAMA, using the never/infrequent wheezing as the baseline category. Analyses were carried out in SPSS using a logistic regression model.

Appendix 5—table 1
Single nucleotide polymorphisms (SNPs) near ANXA1 associated with persistent wheeze.
ChrRsidPositionA1A2freqA2BetaSEp-ValueDirection(3 GWAS)
9rs7526065475788108tc0.020.900.166.66e-09---
9rs11684966475820902tc0.020.890.161.99e-08---
9rs7832098475844302tg0.020.810.156.41e-08---

  1. A1 is the effect allele, A2 is the reference allele.

Appendix 5—table 2
Allele frequencies of rs75260654 across different wheeze phenotypes.
PhenotypeCCCTTT
Never/infrequent5641 (97.2)161 (2.8)0 (0)
Early-onset pre-school remitting1409 (97.1)42 (2.9)0 (0)
Early-onset mid-childhood remitting572 (96.1)23 (3.9)0 (0)
Late-onset613 (95.2)31 (4.8)0 (0)
Early-onset persistent867 (94.2)51 (5.5)2 (0.2)
Appendix 5—table 3
Selected immune eQTLs of rs75260654.
Rsidp-ValueBetaSEanSymbolStudy
rs752606540.014–0.650.26382ANXA1Quach_2016_monocyte_R848
rs752606540.015–1.020.41396ANXA1Quach_2016_monocyte_IAV
Appendix 5—table 4
Lung eQTLs of rs75260654.
Rsidp-ValueBetaSEanSymbolStudy
rs1168496640.04890.220.11620ANXA1GTEx_exon_lung
rs783209840.04890.220.11620ANXA1GTEx_exon_lung
Appendix 5—table 5
Functional enrichment for ANXA1: top 10 GO terms.
Term nameDescriptionFDR value
GO.0007186G protein-coupled receptor signalling pathway3.57×10–18
GO.0006954Inflammatory response1.13×10–16
GO.0006874Cellular calcium ion homeostasis5.55×10–15
GO.0007204Positive regulation of cytosolic calcium ion concentration1.65×10–14
GO.0060326Cell chemotaxis1.95×10–14
GO.0006955Immune response4.23×10–14
GO.0006935Chemotaxis4.93×10–14
GO.0006952Defense response1.68×10–13
GO.0050801Ion homeostasis2.23×10–13
GO.0002376Immune system process2.87×10–13
Appendix 5—table 6
Replication of associations between single nucleotide polymorphisms (SNPs) downstream of ANXA1 and early-onset persistent wheezing in Prevention and Incidence of Asthma and Mite Allergy (PIAMA).
CW (40) vs NI (1557)
Rsidchr:positionR2A2/freqA2BetaSEp-Value
rs752606549:757881080.60c/0.02–0.2870.910.75
rs1168496649:758209020.61c/0.020.1191.080.91
rs783209849:758443020.59g/0.020.1251.040.90

  1. A2 is the reference allele. CW = continuous wheezing, IR = intermediate wheezing derived from LCA 1–12 years in PIAMA.

Appendix 5—figure 1
Download asset Open asset
Linkage disequilibrium between single nucleotide polymorphisms (SNPs) downstream of ANXA1 that were associated with persistent wheeze.
Appendix 5—figure 2
Download asset Open asset
ANXA1 interactors.

Protein-protein interaction of ANXA1 including IL-4, IL-13, and NR3C1.

Appendix 6

Functional mouse experiments

Mice

In accordance with the Animals (scientific procedures) act 1986, all animal experiments were conducted under the approved UK Home Office Project License No: PPL 70/7643, reviewed by Imperial College’s Animal Welfare and Ethical Review body. Female WT BALB/c and annexin A1 knock-out mice were purchased from Charles River (Bicester, UK). Animals aged 6–8 weeks of age received 25 µg intranasal instillation of either HDM (Greer Laboratories, Lenoir, NC, USA; Cat: XPB70D3A25) or PBS 3× a week for 3 weeks. Mice were sacrificed 4 hr post-final HDM challenge. Mice were housed under specific pathogen-free conditions and a 12:12 light:dark cycle. Food and water were supplied ad libitum. All animal experiments were completed twice, with N=4–6 per group.

Airway hyperresponsiveness

Airway hyperreactivity was measured using Flexivent. Lung resistance was measured in response to increasing doses of methacholine (3–100 mg/ml, Sigma, Poole, UK, Cat: A2251) in tracheotomised anaesthetised mice using an EMMS system (Electro-Medical Measurement Systems, UK).

Flow cytometry analysis

Bronchoalveolar lavage (BAL) was collected. BAL cells were restimulated with ionomycin and phorbol 12-myristate 13-acetate in the presence of brefeldin (Sigma), as previously described (Branchett et al., 2021). Specific antibodies for T1/ST2 staining were purchased from Morwell Diagnostics (Zurich, Switzerland). Cells were also stained for lineage negative cocktail, Ly6G, CD45, CD11b, CD11c, SiglecF. Labelled cells were acquired on a BD Fortessa (BD Biosciences, Oxford, UK) and analysed using FlowJo software (Treestar, Ashland, OR, USA). Details of antibodies used can be found in Appendix 6—table 1.

Appendix 6—table 1
Antibodies used in the flow cytometry analysis.
MoleculeManufacturerIsotypeConjugated dyeClone
T1/ST2Morwell Diagnostics GMBH, SwitzerlandRat IgG1FITCDJ8
CD45e-Bioscience Ltd, Hatfield, UKRat IgG2bPerCP-CY5.530-F11
CD11bBD Biosciences, Oxford, UKRat IgG2be450M1/70
CD11ce-Bioscience Ltd, Hatfield, UKHamster IgG1PerCP-CY5.5N418
Siglec FBD Biosciences, Oxford, UKRat IgG2aPEE50-2440

Analysis of cytokines and chemokines

Murine lung tissue homogenate supernatants were processed as previously described (Branchett et al., 2021). Cytokine levels were analysed by ELISA: IL-4, IL-5 (PharMingen, Oxford, UK), IL-13 Ready-Set-Go kits (eBioscience).

Real-time PCR

Total RNA was extracted from murine lung tissue using an RNeasy Mini Kit (QIAGEN). Total RNA (1 μg) was reverse transcribed into cDNA using a High-Capacity cDNA Reverse Transcription Kit (Life Technologies, UK). Real-time PCRs were performed using TaqMan Gene Expression Master Mix and TaqMan Gene Expression probes, annexin A1, and HPRT (Applied Biosystems). Values were normalised to HPRT and gene expression was analysed using the change-in-threshold 2-∆CT method.

Annexin A1 immunohistochemistry

Paraffin-embedded mouse lung sections were stained with annexin A1 (R&D Systems, MAB3770). Annexin A1 primary antibody was followed by a secondary detection antibody (donkey anti-goat 488, Thermo Fisher, A11055). Annexin A1+ cells were quantified by manual counting under microscope and numbers averaged over four fields, from five biological replicates per group.

Statistical analysis

Data are expressed as median±IQR. Statistical differences between groups were calculated using Mann-Whitney U test, unless otherwise specified. p-Values are indicated in figures.

Data availability

The informed consent obtained from all included participants does not allow the data to be made freely available through any third party maintained public repository. However, data used for this submission can be made available on request to the corresponding cohort Executive. Researchers will need to submit a research proposal to each cohort Executive Committee. Data access will have a cost, for more details re. ALSPAC contact alspac-data@bristol.ac.uk, for any other cohort contact philip.couch@manchester.ac.uk.The ALSPAC website provides information on how to request and access its data (http://www.bristol.ac.uk/alspac/researchers/access/). For queries regarding access of data from MAAS, IoW, SEATON or Ashford please contact Philip Couch (philip.couch@manchester.ac.uk). All code used to analyse the individual level data and all summary data and code used to plot the figures in our manuscript has been deposited in Dryad.

The following data sets were generated
    1. Granell R
    (2023) Dryad Digital Repository
    A meta-analysis of genome-wide association studies of childhood wheezing phenotypes identifies ANXA1 as a susceptibility locus for persistent wheezing (GWAS ANXA1).
    https://doi.org/10.5061/dryad.3r2280gm3

References

    1. Demenais F
    2. Margaritte-Jeannin P
    3. Barnes KC
    4. Cookson WOC
    5. Altmüller J
    6. Ang W
    7. Barr RG
    8. Beaty TH
    9. Becker AB
    10. Beilby J
    11. Bisgaard H
    12. Bjornsdottir US
    13. Bleecker E
    14. Bønnelykke K
    15. Boomsma DI
    16. Bouzigon E
    17. Brightling CE
    18. Brossard M
    19. Brusselle GG
    20. Burchard E
    21. Burkart KM
    22. Bush A
    23. Chan-Yeung M
    24. Chung KF
    25. Couto Alves A
    26. Curtin JA
    27. Custovic A
    28. Daley D
    29. de Jongste JC
    30. Del-Rio-Navarro BE
    31. Donohue KM
    32. Duijts L
    33. Eng C
    34. Eriksson JG
    35. Farrall M
    36. Fedorova Y
    37. Feenstra B
    38. Ferreira MA
    39. Australian Asthma Genetics Consortium (AAGC) collaborators
    40. Freidin MB
    41. Gajdos Z
    42. Gauderman J
    43. Gehring U
    44. Geller F
    45. Genuneit J
    46. Gharib SA
    47. Gilliland F
    48. Granell R
    49. Graves PE
    50. Gudbjartsson DF
    51. Haahtela T
    52. Heckbert SR
    53. Heederik D
    54. Heinrich J
    55. Heliövaara M
    56. Henderson J
    57. Himes BE
    58. Hirose H
    59. Hirschhorn JN
    60. Hofman A
    61. Holt P
    62. Hottenga J
    63. Hudson TJ
    64. Hui J
    65. Imboden M
    66. Ivanov V
    67. Jaddoe VWV
    68. James A
    69. Janson C
    70. Jarvelin M-R
    71. Jarvis D
    72. Jones G
    73. Jonsdottir I
    74. Jousilahti P
    75. Kabesch M
    76. Kähönen M
    77. Kantor DB
    78. Karunas AS
    79. Khusnutdinova E
    80. Koppelman GH
    81. Kozyrskyj AL
    82. Kreiner E
    83. Kubo M
    84. Kumar R
    85. Kumar A
    86. Kuokkanen M
    87. Lahousse L
    88. Laitinen T
    89. Laprise C
    90. Lathrop M
    91. Lau S
    92. Lee Y-A
    93. Lehtimäki T
    94. Letort S
    95. Levin AM
    96. Li G
    97. Liang L
    98. Loehr LR
    99. London SJ
    100. Loth DW
    101. Manichaikul A
    102. Marenholz I
    103. Martinez FJ
    104. Matheson MC
    105. Mathias RA
    106. Matsumoto K
    107. Mbarek H
    108. McArdle WL
    109. Melbye M
    110. Melén E
    111. Meyers D
    112. Michel S
    113. Mohamdi H
    114. Musk AW
    115. Myers RA
    116. Nieuwenhuis MAE
    117. Noguchi E
    118. O’Connor GT
    119. Ogorodova LM
    120. Palmer CD
    121. Palotie A
    122. Park JE
    123. Pennell CE
    124. Pershagen G
    125. Polonikov A
    126. Postma DS
    127. Probst-Hensch N
    128. Puzyrev VP
    129. Raby BA
    130. Raitakari OT
    131. Ramasamy A
    132. Rich SS
    133. Robertson CF
    134. Romieu I
    135. Salam MT
    136. Salomaa V
    137. Schlünssen V
    138. Scott R
    139. Selivanova PA
    140. Sigsgaard T
    141. Simpson A
    142. Siroux V
    143. Smith LJ
    144. Solodilova M
    145. Standl M
    146. Stefansson K
    147. Strachan DP
    148. Stricker BH
    149. Takahashi A
    150. Thompson PJ
    151. Thorleifsson G
    152. Thorsteinsdottir U
    153. Tiesler CMT
    154. Torgerson DG
    155. Tsunoda T
    156. Uitterlinden AG
    157. van der Valk RJP
    158. Vaysse A
    159. Vedantam S
    160. von Berg A
    161. von Mutius E
    162. Vonk JM
    163. Waage J
    164. Wareham NJ
    165. Weiss ST
    166. White WB
    167. Wickman M
    168. Widén E
    169. Willemsen G
    170. Williams LK
    171. Wouters IM
    172. Yang JJ
    173. Zhao JH
    174. Moffatt MF
    175. Ober C
    176. Nicolae DL
    (2018) Multiancestry association study identifies new asthma risk loci that colocalize with immune-cell enhancer marks
    Nature Genetics 50:42–53.
    https://doi.org/10.1038/s41588-017-0014-7

Decision letter

  1. Anurag Agrawal
    Reviewing Editor; Ashoka University, India
  2. Molly Przeworski
    Senior Editor; Columbia University, United States
  3. Anurag Agrawal
    Reviewer; Ashoka University, India

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "A meta-analysis of genome-wide association studies of childhood wheezing phenotypes identifies ANXA1 as a susceptibility locus for persistent wheezing" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Anurag Agrawal as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Molly Przeworski as the Senior Editor.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) Please provide more details on the meta-analysis strategy and rationale

Consider the other comments for greater clarity of the work

Reviewer #1 (Recommendations for the authors):

The work is interesting, novel, and technically rigorous. Some suggestions follow regarding the experimental mouse model part of the manuscript, which needs further editing for clarity and standardisation of terminology.

1. Figure 4A, font size needs to be increased.

2. Line 289-291 needs correction. 4E shows methacholine-induced bronchoconstriction, not allergen-induced reduced lung function.

3. Was baseline resistance (pre-methacholine) different? Any differences in the slope of the curve?

4. Airway remodelling changes should be described if possible.

Reviewer #2 (Recommendations for the authors):

Even though the study described by Granell et al. is an impressive piece of work with robust sets of evidence that support the derived conclusions, this reviewer has some suggestions the authors might want to consider:

1. A more detailed description of the association models applied to test the association with wheezing phenotypes, including the covariates used, is encouraged.

2. Which approach was used for the meta-analysis? Did you take into account the heterogeneity among studies to choose fixed or random-effect models to be applied?

3. Could you please provide an explanation of the genomic inflation found for the persistent wheezing phenotype (Figure E1)?

4. Table E4: Could you please explain the reason for inclusion/exclusion?

Reviewer #3 (Recommendations for the authors):

1. Regarding the definition of wheeze phenotypes from infancy to adolescence, were transitions across different phenotypes over time possible? Was time-varying phenotype of concern? Did the authors evaluate the accuracy of phenotyping using metrics such as sensitivity, specificity, and positive and negative predictive values? And how much of the variation in each wheeze phenotype can be explained by genetic profile?

2. In order to identify primary and secondary independent loci, can the authors perform the conditional analysis? Further, considering the LD structure, credible set analysis, and colocalization analysis can help identify potential causal variants.

3. Can the authors provide further explanations on the biology using pQTL and TWAS?

4. Can the authors use DEPICT or other tools to identify relevant tissues/cell types that are highly expressed?

5. The heterogeneity in the genetic profile for the wheeze phenotypes can be better illustrated, such as by using a scatter plot to show the differences in genetic effect sizes. A heterogeneity p-value can be computed. In addition, what was the genetic correlation between the wheeze phenotypes?

https://doi.org/10.7554/eLife.84315.sa1

Author response

Essential revisions:

1) Please provide more details on the meta-analysis strategy and rationale

Please see Reviewer 2 Q2.

Consider the other comments for greater clarity of the work

Reviewer #1 (Recommendations for the authors):

The work is interesting, novel, and technically rigorous. Some suggestions follow regarding the experimental mouse model part of the manuscript, which needs further editing for clarity and standardisation of terminology.

1. Figure 4A, font size needs to be increased.

Figure 5A (4Ain previous version) has been updated.

2. Line 289-291 needs correction. 4E shows methacholine-induced bronchoconstriction, not allergen-induced reduced lung function.

We have corrected this in line 322 of the Revised Manuscript.

3. Was baseline resistance (pre-methacholine) different? Any differences in the slope of the curve?

No statistically significant difference was observed in resistance at baseline. However, a leftward shift in the curve was observed.

4. Airway remodelling changes should be described if possible.

We did not assess changes in remodelling in these experiments. Here, we utilised a 3-week HDM challenge model, whereas to fully characterise changes in airways remodelling, mice would need to be assessed later than 3 weeks post HDM.

Reviewer #2 (Recommendations for the authors):

Even though the study described by Granell et al. is an impressive piece of work with robust sets of evidence that support the derived conclusions, this reviewer has some suggestions the authors might want to consider:

1. A more detailed description of the association models applied to test the association with wheezing phenotypes, including the covariates used, is encouraged.

We agree with this reviewer’s comment, and we have made this clearer in the Methods section on lines 185-7 Revised Manuscript:

“We used SNPTEST v2.5.2(26) with a frequentist additive multinomial logistic regression model (-method newml), using the never/infrequent wheeze as the reference and without including any covariates.”

2. Which approach was used for the meta-analysis? Did you take into account the heterogeneity among studies to choose fixed or random-effect models to be applied?

These details were already available in the Supplement, and they can now be found in Appendix 2: GWAS Meta-analysis:

“We used the option SCHEME STDERR in METAL to implement an effect-size based method weighted by each study-specific standard error in a fixed-effects model.”

3. Could you please provide an explanation of the genomic inflation found for the persistent wheezing phenotype (Figure E1)?

We would interpret Figure 1—figure supplement 1 (previously Figure E1) as an association between SNPs and Persistent wheeze rather than inflation. For instance, we did not observe any deviation in the other phenotypes. We now report the phenotype specific genomic inflation factor for each wheezing phenotype in lines 222-5 Revised Manuscript:

“We observed slight deflation of the meta-analysis pvalues in our summary statistics. Genomic Inflation Factor (λ) for Early-onset Pre-school Remitting=0.96, Early-onset Mid-childhood Remitting = 0.94, Late-onset=0.96, and Early-onset Persistent wheezing=0.97.”

And have added this text in Appendix 2: Genetic Control:

“The genomic inflation factor (λ) was calculated using the scipy.stats.chi2 module in Python. The chi-squared test statistics from the meta-analysis p-values were first obtained. Then the observed median chi-squared statistic from the calculated chi-squared test statistics were calculated. Finally, the genomic inflation factor (λ) was derived by dividing the observed median chi-squared statistic by the expected median chi-squared statistic.”

4. Table E4: Could you please explain the reason for inclusion/exclusion?

Thank you, this was not properly explained in the main text. Individuals with missing genetic data, as well as related and non-European individuals were excluded. We have now made this clear in line 212-4 Revised Manuscript:

“Individuals with missing genetic data, as well as related and non-European individuals were excluded. Comparison of included vs. excluded individuals across cohorts (per cohort and time point) is shown in Appendix 1 and Appendix 1-Table 4.”

Reviewer #3 (Recommendations for the authors):

1. Regarding the definition of wheeze phenotypes from infancy to adolescence, were transitions across different phenotypes over time possible? Was time-varying phenotype of concern? Did the authors evaluate the accuracy of phenotyping using metrics such as sensitivity, specificity, and positive and negative predictive values?

No, LCA does not allow transitions across different phenotypes over time.

Was time-varying phenotype of concern? Did the authors evaluate the accuracy of phenotyping using metrics such as sensitivity, specificity, and positive and negative predictive values?

We have explored time-varying phenotype in the context of LCA in our recent paper (Haider S et al. Modeling Wheezing Spells Identifies Phenotypes with Different Outcomes and Genetic Associates. Am J Respir Crit Care Med. 2022 Apr 15;205(8):883-893. doi: 10.1164/rccm.202108-1821OC. PMID: 35050846; PMCID: PMC9838626.) where we compared the robustness of LCA and PAM phenotypes. In this paper we show that latent class analysis- and Partition Around Medoids (PAM) clustering of spell-based phenotypes appeared similar, but within-phenotype individual trajectories and phenotype allocation differed substantially. The spell-based approach was much more robust in dealing with missing data, and the derived clusters were more stable and internally homogeneous than those derived by LCA. If anything, any instability of LCA phenotypes over time would reduce the power of the current study, yet we still identified genome-wide significant associations unique to specific wheeze clusters. None-the-less future studies should explore the utilisation of PAM derived phenotypes in genetic analysis of allergic disease.

And how much of the variation in each wheeze phenotype can be explained by genetic profile?

We have not explored this question in this paper, but we will do so in our follow-up paper.

2. In order to identify primary and secondary independent loci, can the authors perform the conditional analysis? Further, considering the LD structure, credible set analysis, and colocalization analysis can help identify potential causal variants.

3. Can the authors provide further explanations on the biology using pQTL and TWAS?

4. Can the authors use DEPICT or other tools to identify relevant tissues/cell types that are highly expressed?

The reviewer makes a number of helpful suggestions of additional post-GWAS analyses that could be undertaken on our results. Given the size of the manuscript and the extensive incorporation of mouse model data for the ANAX1 loci, we feel that adding further post GWAS analyses to the supplementary data is beyond the scope of this paper and we are currently working on a new follow-up paper that will focus on fine mapping of these loci.

5. The heterogeneity in the genetic profile for the wheeze phenotypes can be better illustrated, such as by using a scatter plot to show the differences in genetic effect sizes. A heterogeneity p-value can be computed.

We agree with the reviewer’s suggestion and have produced scatter plots to illustrate the heterogeneity in the genetic profile of the wheezing phenotypes. We have added these plots in Figure 2—figure supplement 1 and the following text has been added in lines 230-3 Revised Manuscript:

“Scatter plots in Figure 2—figure supplement 1 show the heterogeneity in the genetic profile of the wheeze phenotypes. The plots show that all signals were phenotype-specific at p<10-5 and only nominal associations were shared across wheezing phenotypes. More details on how these plots were derived can be found in Appendix 2: Heterogeneity scatter plots.”

In addition, what was the genetic correlation between the wheeze phenotypes?

We identified exclusive subgroups of SNPs associated with each wheezing phenotype, and only two locus were shared with Persistent and Early-onset Mid-childhood Remitting Wheezing (chr17q12 and chr17q21) therefore the genetic correlation between the wheeze phenotypes is likely to be small.

https://doi.org/10.7554/eLife.84315.sa2

Article and author information

Author details

  1. Raquel Granell

    MRC Integrative Epidemiology Unit, Department of Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, United Kingdom
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Methodology, Writing – original draft, Writing – review and editing
    For correspondence
    Raquel.Granell@bristol.ac.uk
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4890-4012

  2. John A Curtin

    Division of Infection, Immunity and Respiratory Medicine, School of Biological Sciences, The University of Manchester, Manchester Academic Health Science Centre, and Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Contribution
    Formal analysis, Visualization, Methodology, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared

  3. Sadia Haider

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Formal analysis, Writing – review and editing
    Competing interests
    No competing interests declared

  4. Negusse Tadesse Kitaba

    Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    Contribution
    Writing – review and editing, Contributed to interpretation of results, Post-gwas analyses
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7518-9096

  5. Sara A Mathie

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Writing – review and editing, Contributed to work related to mouse models
    Competing interests
    No competing interests declared

  6. Lisa G Gregory

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Writing – original draft, Writing – review and editing, Contributed to work related to mouse models
    Competing interests
    No competing interests declared

  7. Laura L Yates

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Writing – review and editing, Contributed to work related to mouse models
    Competing interests
    No competing interests declared

  8. Mauro Tutino

    Division of Infection, Immunity and Respiratory Medicine, School of Biological Sciences, The University of Manchester, Manchester Academic Health Science Centre, and Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Contribution
    Visualization, Writing – review and editing, Contributed to interpretation of results post-gwas
    Competing interests
    No competing interests declared

  9. Jenny Hankinson

    Division of Infection, Immunity and Respiratory Medicine, School of Biological Sciences, The University of Manchester, Manchester Academic Health Science Centre, and Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Contribution
    Writing – review and editing, Designed and carried out the HiC work
    Competing interests
    No competing interests declared

  10. Mauro Perretti

    William Harvey Research Institute, Barts and The London School of Medicine Queen Mary University of London, London, United Kingdom
    Contribution
    Writing – review and editing, Contributed to work related to mouse models
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-2068-3331

  11. Judith M Vonk

    1. Department of Epidemiology, University of Groningen, University Medical Center Groningen\, Groningen, Netherlands
    2. University of Groningen, University Medical Center Groningen, Groningen Research Institute for Asthma and COPD (GRIAC), Groningen, Netherlands
    Contribution
    Writing – review and editing, Replication in PIAMA
    Competing interests
    No competing interests declared

  12. Hasan S Arshad

    1. NIHR Southampton Biomedical Research Centre, University Hospitals Southampton NHS Foundation Trust, Southampton, United Kingdom
    2. David Hide Asthma and Allergy Research Centre, Isle of Wight, United Kingdom
    3. Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    No competing interests declared

  13. Paul Cullinan

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    No competing interests declared

  14. Sara Fontanella

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    No competing interests declared

  15. Graham C Roberts

    1. Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospitals Southampton NHS Foundation Trust, Southampton, United Kingdom
    3. David Hide Asthma and Allergy Research Centre, Isle of Wight, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    MRC grant to my institution; President of the British Society of Allergy and Clinical Immunology

  16. Gerard H Koppelman

    1. University of Groningen, University Medical Center Groningen, Groningen Research Institute for Asthma and COPD (GRIAC), Groningen, Netherlands
    2. Department of Pediatric Pulmonology and Pediatric Allergology, University of Groningen, University Medical Center Groningen, Beatrix Children’s Hospital, Groningen, Netherlands
    Contribution
    Writing – review and editing, Replication in PIAMA
    Competing interests
    Dutch Lung Foundation, Ubbo Emmius Foundation (Money to insititution); Dutch Lung Foundation, Vertex, TEVA the Netherlands, GSK, ZON-MW (VICI grant), European Union (Money to institution); Astra Zeneca, Pure IMS, GSK (Money to institution); Sanofi, Boehringer Ingelheim (Money to institution)

  17. Angela Simpson

    Division of Infection, Immunity and Respiratory Medicine, School of Biological Sciences, The University of Manchester, Manchester Academic Health Science Centre, and Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    Medical research council Research grant; JP Moulton Charitable Foundation Research grant; Asthma UK Research grant

  18. Steve W Turner

    Child Health, University of Aberdeen, Aberdeen, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    No competing interests declared

  19. Clare S Murray

    Division of Infection, Immunity and Respiratory Medicine, School of Biological Sciences, The University of Manchester, Manchester Academic Health Science Centre, and Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Contribution
    Writing – review and editing
    Competing interests
    has received grants from Asthma Uk, the National Institute for Health Research, the Moulton Charitable Foundation and the North West Lung Centre Charity (to the Institution). They received lecture fees from GSK and Novartis, and received a travel grant from Sanofi. The authors has no other competing interests to declare

  20. Clare M Lloyd

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Writing – review and editing, Contributed to work related to mouse models
    Competing interests
    Wellcome Trust 107059/Z/15/Z

  21. John W Holloway

    1. Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, United Kingdom
    2. NIHR Southampton Biomedical Research Centre, University Hospitals Southampton NHS Foundation Trust, Southampton, United Kingdom
    Contribution
    Conceptualization, Writing – review and editing
    Competing interests
    Medial Research Council grant MR/S025340/1 (to institution); American Academy of Allergy Asthma and Immunology (AAAI) (Support for speaker travel to AAAAI annual congress)
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9998-0464

  22. Adnan Custovic

    National Heart and Lung Institute, Imperial College London, London, United Kingdom
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Writing – original draft
    For correspondence
    a.custovic@imperial.ac.uk
    Competing interests
    MRC (research grants); EPSRC (research grant); Wellcome Trust (research grant); Worg Pharmaceoticals (Personal payment)
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5218-7071

Funding

UK Medical Research Council (MR/S025340/1)

  • Raquel Granell
  • Adnan Custovic

Wellcome Trust (108818/15/Z)

  • Raquel Granell
  • Adnan Custovic

Wellcome Trust (107059/Z/15/Z)

  • Clare M Lloyd

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. For the purpose of Open Access, the authors have applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

Acknowledgements

STELAR/UNICORN investigators: Professor Graham Devereux, Dr Dimitrios Charalampopoulos. Breathing Together investigators: Prof Andrew Bush, Prof Sejal Saglani, Prof Benjamin Marsland, Prof Jonathan Grigg, Prof Jurgen Schwarze, Prof Mike Shields, Prof Peter Ghazal, Prof Ultan Power, Dr Ceyda Oksel.

Supported by the UK Medical Research Council (MRC) Programme Grant MR/S025340/1, the Wellcome Trust (WT) Strategic Award (108818/15/Z) and a WT Senior Fellowship to CML (107059/Z/15/Z). The MRC and Wellcome (Grant ref: 217065/Z/19/Z) and the University of Bristol provide core support for ALSPAC. ALSPAC GWAS data was generated by Sample Logistics and Genotyping Facilities at Wellcome Sanger Institute and LabCorp (Laboratory Corporation of America) using support from 23andMe. A comprehensive list of grants funding is available on the ALSPAC website (http://www.bristol.ac.uk/alspac/external/documents/grant-acknowledgements.pdf).

PIAMA was funded by the Netherlands Lung Foundation (grant 3.4.01.26, 3.2.06.022, 3.4.09.081 and 3.2.10.085CO), the ZON-MW Netherlands Organization for Health Research and Development (grant 912-03-031), the Stichting Astmabestrijding and the Ministry of the Environment. Genome-wide genotyping was funded by the European Commission as part of GABRIEL (grant number 018996) and a grant from BBMRI-NL (CP 29). GHK is supported by a ZON-MW VICI grant.

For the purpose of Open Access, the author has applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

Ethics

ALSPAC: Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committees. Informed consent for the use of data collected via questionnaires and clinics was obtained from participants following the recommendations of the ALSPAC Ethics and Law Committee at the time. All self-completion questionnaire content is approved by the ALSPAC Ethics and Law Committee. Bristol and Weston Health Authority: E1808 Children of the Nineties: Avon Longitudinal Study of Pregnancy and Childhood (ALSPAC). (28th November 1989); Southmead Health Authority: 49/89 Children of the Nineties -"ALSPAC". (5th April 1990); Frenchay Health Authority: 90/8 Children of the Nineties. (28th June 1990).MAAS: The study was approved by the North West - Greater Manchester East Research Ethics Committee. ERP/94/032 Up to 5 yrs. Allergen avoidance, Primary Prevention, genetics, sRaw age 3 and 5; SOU/00/259 5 year; ERP/95/137 Exposure to pet allergens, atopy, genetics; ERP/97/023 IFWIN, genetics; 03/SM/400 8 year; 06/Q1403/142 10-12 years; 11/NW/0228 13-15 years; 14/NW/1309 18+ years.SEATON: The study was approved by the North of Scotland Research Ethics Committee. REC reference: 13/NS/0108; Protocol number: 2/048/13; Amendment number: AM03.Ashford: The Asthma in Ashford study was reviewed by the Imperial College Research Ethics Committee on 11/11/2014. On 08/01/2015 the Joint Research Compliance Office granted full approval of the study on the basis described in the revised documents. ICREC reference: 14|C2288.IOW: Ethics approval for the IoW cohort was originally given by the Isle of Wight local research ethics committee in 1989 and at each subsequent follow up (1,2 and 4 years) (this is pre "numbers")Age 10 follow up (including DNA and genotyping): Isle of Wight Health Authority Local Research Ethics Committee 18/98. Age 18 Follow up(including DNA and genotyping): Isle of Wight, Portsmouth & South East Hampshire Research Ethics Committee 06/Q1701/34.

In accordance with the Animals (scientific procedures) act 1986, all animal experiments were conducted under the approved UK Home Office Project License No: PPL 70/7643, reviewed by Imperial College's Animal Welfare and Ethical Review body.

Senior Editor

  1. Molly Przeworski, Columbia University, United States

Reviewing Editor

  1. Anurag Agrawal, Ashoka University, India

Reviewer

  1. Anurag Agrawal, Ashoka University, India

Version history

  1. Received: October 19, 2022
  2. Preprint posted: March 5, 2023 (view preprint)
  3. Accepted: May 22, 2023
  4. Accepted Manuscript published: May 25, 2023 (version 1)
  5. Version of Record published: June 26, 2023 (version 2)

Copyright

© 2023, Granell et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 657
    Page views
  • 81
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Raquel Granell
  2. John A Curtin
  3. Sadia Haider
  4. Negusse Tadesse Kitaba
  5. Sara A Mathie
  6. Lisa G Gregory
  7. Laura L Yates
  8. Mauro Tutino
  9. Jenny Hankinson
  10. Mauro Perretti
  11. Judith M Vonk
  12. Hasan S Arshad
  13. Paul Cullinan
  14. Sara Fontanella
  15. Graham C Roberts
  16. Gerard H Koppelman
  17. Angela Simpson
  18. Steve W Turner
  19. Clare S Murray
  20. Clare M Lloyd
  21. John W Holloway
  22. Adnan Custovic
  23. on behalf of UNICORN and Breathing Together investigators
(2023)
A meta-analysis of genome-wide association studies of childhood wheezing phenotypes identifies ANXA1 as a susceptibility locus for persistent wheezing
eLife 12:e84315.
https://doi.org/10.7554/eLife.84315

Categories and tags

Research organism

Further reading

    1. Biochemistry and Chemical Biology
    2. Epidemiology and Global Health

    Status and physiological significance of circulating adiponectin in the very old and centenarians: an observational study

    Takashi Sasaki, Yoshinori Nishimoto ... Yasumichi Arai
    Research Article

    Background: High levels of circulating adiponectin are associated with increased insulin sensitivity, low prevalence of diabetes, and low body mass index (BMI); however, high levels of circulating adiponectin are also associated with increased mortality in the 60-70 age group. In this study, we aimed to clarify factors associated with circulating high-molecular-weight (cHMW) adiponectin levels and their association with mortality in the very old (85-89 years old) and centenarians.

    Methods: The study included 812 (women: 84.4%) for centenarians and 1,498 (women: 51.7%) for the very old. The genomic DNA sequence data were obtained by whole genome sequencing or DNA microarray-imputation methods. LASSO and multivariate regression analyses were used to evaluate cHMW adiponectin characteristics and associated factors. All-cause mortality was analyzed in three quantile groups of cHMW adiponectin levels using Cox regression.

    Results: The cHMW adiponectin levels were increased significantly beyond 100 years of age, were negatively associated with diabetes prevalence, and were associated with SNVs in CDH13 (p = 2.21 × 10-22) and ADIPOQ (p = 5.72 × 10-7). Multivariate regression analysis revealed that genetic variants, BMI, and high-density lipoprotein cholesterol (HDLC) were the main factors associated with cHMW adiponectin levels in the very old, whereas the BMI showed no association in centenarians. The hazard ratios for all-cause mortality in the intermediate and high cHMW adiponectin groups in very old men were significantly higher rather than those for all-cause mortality in the low level cHMW adiponectin group, even after adjustment with BMI. In contrast, the hazard ratios for all-cause mortality were significantly higher for high cHMW adiponectin groups in very old women, but were not significant after adjustment with BMI.

    Conclusions: cHMW adiponectin levels increased with age until centenarians, and the contribution of known major factors associated with cHMW adiponectin levels, including BMI and HDLC, varies with age, suggesting that its physiological significance also varies with age in the oldest old.

    Funding: This study was supported by grants from the Ministry of Health, Welfare, and Labour for the Scientific Research Projects for Longevity; a Grant-in-Aid for Scientific Research (No 21590775, 24590898, 15KT0009, 18H03055, 20K20409, 20K07792, 23H03337) from the Japan Society for the Promotion of Science; Keio University Global Research Institute (KGRI), Kanagawa Institute of Industrial Science and Technology (KISTEC), Japan Science and Technology Agency (JST) Research Complex Program 'Tonomachi Research Complex' Wellbeing Research Campus: Creating new values through technological and social innovation (JP15667051), the Program for an Integrated Database of Clinical and Genomic Information from the Japan Agency for Medical Research and Development (No. 16kk0205009h001, 17jm0210051h0001, 19dk0207045h0001); the medical-welfare-food-agriculture collaborative consortium project from the Japan Ministry of Agriculture, Forestry, and Fisheries; and the Biobank Japan Program from the Ministry of Education, Culture, Sports, and Technology.

    1. Epidemiology and Global Health

    Prediction of diabetic kidney disease risk using machine learning models: A population-based cohort study of Asian adults

    Charumathi Sabanayagam, Feng He ... Ching Yu Cheng
    Research Article Updated

    Background:

    Machine learning (ML) techniques improve disease prediction by identifying the most relevant features in multidimensional data. We compared the accuracy of ML algorithms for predicting incident diabetic kidney disease (DKD).

    Methods:

    We utilized longitudinal data from 1365 Chinese, Malay, and Indian participants aged 40–80 y with diabetes but free of DKD who participated in the baseline and 6-year follow-up visit of the Singapore Epidemiology of Eye Diseases Study (2004–2017). Incident DKD (11.9%) was defined as an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2 with at least 25% decrease in eGFR at follow-up from baseline. A total of 339 features, including participant characteristics, retinal imaging, and genetic and blood metabolites, were used as predictors. Performances of several ML models were compared to each other and to logistic regression (LR) model based on established features of DKD (age, sex, ethnicity, duration of diabetes, systolic blood pressure, HbA1c, and body mass index) using area under the receiver operating characteristic curve (AUC).

    Results:

    ML model Elastic Net (EN) had the best AUC (95% CI) of 0.851 (0.847–0.856), which was 7.0% relatively higher than by LR 0.795 (0.790–0.801). Sensitivity and specificity of EN were 88.2 and 65.9% vs. 73.0 and 72.8% by LR. The top 15 predictors included age, ethnicity, antidiabetic medication, hypertension, diabetic retinopathy, systolic blood pressure, HbA1c, eGFR, and metabolites related to lipids, lipoproteins, fatty acids, and ketone bodies.

    Conclusions:

    Our results showed that ML, together with feature selection, improves prediction accuracy of DKD risk in an asymptomatic stable population and identifies novel risk factors, including metabolites.

    Funding:

    This study was supported by the National Medical Research Council, NMRC/OFLCG/001/2017 and NMRC/HCSAINV/MOH-001019-00. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

    1. Epidemiology and Global Health

    The importance of patient-partnered research in addressing long COVID: Takeaways for biomedical research study design from the RECOVER Initiative’s Mechanistic Pathways taskforce

    C Kim, Benjamin Chen ... RECOVER Mechanistic Pathways Task Force
    Review Article

    The NIH-funded RECOVER study is collecting clinical data on patients who experience a SARS-CoV-2 infection. As patient representatives of the RECOVER Initiative’s Mechanistic Pathways task force, we offer our perspectives on patient motivations for partnering with researchers to obtain results from mechanistic studies. We emphasize the challenges of balancing urgency with scientific rigor. We recognize the importance of such partnerships in addressing post-acute sequelae of SARS-CoV-2 infection (PASC), which includes ‘long COVID,’ through contrasting objective and subjective narratives. Long COVID’s prevalence served as a call to action for patients like us to become actively involved in efforts to understand our condition. Patient-centered and patient-partnered research informs the balance between urgency and robust mechanistic research. Results from collaborating on protocol design, diverse patient inclusion, and awareness of community concerns establish a new precedent in biomedical research study design. With a public health matter as pressing as the long-term complications that can emerge after SARS-CoV-2 infection, considerate and equitable stakeholder involvement is essential to guiding seminal research. Discussions in the RECOVER Mechanistic Pathways task force gave rise to this commentary as well as other review articles on the current scientific understanding of PASC mechanisms.