Childhood asthma is a heterogeneous disease. Asthma classification is typically based on endotypes, which are defined as subtypes of a disease based on shared pathophysiological mechanisms. The introduction of novel treatments with biologics targeting type 2 inflammation pathways urges the development of clinical decision-making tools to guide therapy based on underlying asthma endotypes driving the disease in an individual patient. Access to novel treatment options with biologicals is currently quite random and relies on very few and crude indicators. A personalized medicine approach may benefit the patient as unnecessary treatments are avoided by better matching of patients and therapies. In addition, our current understanding of asthma endotypes is limited and most asthma endotypes involve concomitant inflammatory pathways with dynamic interactions between those pathways that may or may not present in all patients, or in each patient at all time points . For these reasons, omics approaches might provide novel insights in asthma endotypes [2▪]. This is especially important for pediatric asthma, since data in this population are scarce. The recently published ERS/ATS Task Force report on the management of severe asthma  stresses the importance of performing pediatric trials on novel biologicals in children, as well as studies on biomarkers guiding these treatments.
In this review, we highlight the latest developments in various asthma omics fields (genomics, epigenomics, transcriptomics, metabolomics, and proteomics), and describe how these have contributed to our current understanding of childhood asthma endotypes, taking into account challenges and opportunities of the different approaches. (Table 1) Lastly, we will discuss the most recent efforts towards integrating different omics approaches and how this could address clinical needs, such as guiding treatment.
One of strongest genetic risk predictors for childhood asthma is the 17q12–21 region , but it remains unclear which genes are causing this increased risk . Prime suspects are ORMDL3 and GSDMB, but also GSDMA and PGAP3 have been named as potential candidates. There are several other genes strongly associated with asthma, of which IL33, IL1RL1 (encoding the IL-33 receptor, also known as ST2), and thymic stromal lymphopoetin have emerged as promising targets for novel asthma treatments. In addition, a large recent meta-analysis of genome-wide association studies (including almost 24 000 asthma cases, >118 500 healthy controls) identified nine novel loci (adding up 18 loci in total), the majority involved in the immune response to viruses or bacteria , underscoring the importance of host defense processes in asthma risk. Although genomics studies have provided more insights into the genetic architecture of asthma, the contribution of single genetic variants to the asthma risk is low . The integration of environmental exposures in genome-wide interaction studies might provide additional risk information and explain part of the missing asthma heritability [8,9].
Various genetic variants have been reported to influence corticosteroid response; however, the results remain largely inconsistent and effect sizes are small, suggesting that a genomic susceptibility alone is not enough to drive asthma treatment outcomes [10▪▪]. Remarkably, the strongest predictor for childhood asthma onset, the 17q12–21 region, has recently also been associated with an increased risk of asthma exacerbations  and worse lung function improvement in children treated with inhaled corticosteroids (ICS) . These data suggest, that genetic variation in the 17q12–21 region not only increases asthma risk, but also alters a response to treatment. Closer to clinical implementation, is the pharmacogenomics of long-acting beta-2 agonists (LABAs). A large meta-analysis showed that a genetic variant in the ADRB2 gene (encoding the beta-2 adrenergic receptor) influences the risk of poor response to LABAs in children ; 52% increased risk for exacerbation per risk allele when treated with add on LABAs. This variant (rs1042713, Arg16Gly) is relatively common, and a randomized control trial to guide treatment based on this variant is currently ongoing . Such biomarker-guided trial is key to assess the clinical value of adapting treatment-based individual risk profile and move towards a precision medicine approach.
EPIGENOMICS AND TRANSCRIPTOMICS
Since the epigenome and transcriptome is modified by environmental factors, epigenomic and transcriptomic profiling might provide added value for individual prediction models of asthma outcomes in addition to genomic profiling .
A meta-analysis of EWAS (epigenome-wide association studies) [16▪] showed that reduced DNA methylation of specific CpG sites (in genes involved in the activation of eosinophils and cytotoxic T cells) was associated with an increased risk of childhood asthma. However, in cross-sectional studies, it remains unclear whether these observed differences are driving the disease or are effects of the disease. Novel loci differentially methylated in newborns have been demonstrated as potential biomarkers of risk of asthma by school age. At the same time, cross-sectional associations (asthmatic cases/controls) of methylation patterns in children were shown to reflect both asthma risk as effects of having asthma [17▪▪]. A recent EWAS in nasal swabs collected in children found distinct methylation patterns to be associated with asthma onset and asthma characteristics (elevated IgE/FeNO) in genes that may alter the structure and function of epithelial cells . In addition to asthma onset, methylation patterns in peripheral blood have also been associated with ICS treatment response in asthmatic children . However, ICS use per se does not seem to have any significant influence on blood methylation profiles .
Transcriptomic studies in asthmatic children are scarce. A recent study demonstrated that low gene expression of type I IFN in nasal samples at baseline predicts short-term exacerbation risk in children and, combined with high type 2 inflammation at baseline, creates a specific ‘at-risk’ immune state, when a child becomes highly susceptible to exacerbations [21▪]. A microarray analysis of immortalized B cells from ICS treated children who were part of a clinical trial using a gene regulatory network approach, showed that cultured B cells from poor responders to ICS had increased antiapoptotic pathway regulation, compared with good responders . A U-BIOPRED study in adults identified five distinct transcriptomic signatures based on nasal brushes when comparing childhood onset with adult onset asthma, which might indicate different underlying pathways  and it would be interesting to assess whether similar signatures can be identified in asthmatic children.
Transcriptomics studies are often performed in heterogeneous samples (constituting of multiple-cell types), complicating the interpretation of the results. Yet, recent findings from collaborative American and European efforts suggest that transcriptomics data from whole blood cells (including key genes such as IL1RL1, IL5, and IL17B) could be used to develop predictive models of atopic asthma . These results warrant further prospective studies incorporating expression data, in conjunction with clinical and other biomarker data, into asthma prediction models [25▪].
Furthermore, single-cell omics approaches (e.g., single-cell RNA sequencing) are emerging . This approach allows for gene expression profiling at single-cell level. Although these approaches are still very costly, it is expected that they will become more widely used in the coming years.
METABOLOMICS OF EXHALED BREATH
The analysis of the metabolic content of exhaled breath has received much attention in the past years, since it is a noninvasive method with the potential to provide valuable clinical information on processes in lower airways. Breath contains volatile organic compounds (VOCs) that can be of exogenous (bacteria) or endogenous (cellular) origin. VOC detection can be based on individual compound detection or recognition of patterns of VOC mixtures. Although other techniques exist, most studies in the field of childhood asthma have applied gas chromatography–mass spectrometry (GC–MS) to study individual VOCs or eNOSE technology to study patterns of VOC mixtures. GC–MS is still considered the gold standard since it enables identification of compounds that allows to gain insight in pathophysiological processes and to validate the origin of detected compounds. However, its main limitations are the need for highly trained research personnel and the fact that no real-time (online) findings can be obtained [27▪▪]. Electronic noses, on the other hand, are fast, cheap and easy to use and therefore an appealing point of care technology [27▪▪]. Previous studies have shown that VOC measurements can distinguish inflammatory phenotypes of adult patients with chronic respiratory disease , can predict loss-of-asthma control in adults [29▪] and preliminary data show that VOCs can even predict efficacy of mepolizumab treatment in adult asthmatics . However, very limited data are available in asthmatic children [27▪▪]. VOCs analysis using GC–MS is able to successfully predict exacerbations in asthmatic children [31,32]. More research is needed to validate VOC analysis for clinical use in pediatric asthma and assess the diagnostic value of a point-of-care eNose test.
Systems-wide proteomic tools contribute to a better understanding of a (pato)physiological state of a cell and of a local tissue microenvironment. Proteomic data offer a great addition to transcriptomics and genomics asthma studies, given the complex regulation of gene translation that may be influenced by preexisting asthma pathology, abundance of alternative splicing gene variants, and posttranslational modification (PTM) of synthesized proteins, including histone PTM and phosphoproteomics. Several methods are used for quantitative and qualitative measurements of protein expression and production in biological samples, however, nano HPLC-coupled high resolution MS (nHPLC-HRMS) and immunoassays (flow cytometry, ELISA, Western blotting, and immunohistochemistry) are among the most commonly used. Proteomics is a promising tool for sub-phenotyping asthma patients, identification of biomarkers that allows the estimation of a disease progression and severity, or response to therapy. A recent study in adults identified 10 clusters (or proteotypes) with distinct proteomic signatures within the U-BIOPRED cohort subjects. After overlaying sputum granulocyte counts onto the 10 clusters as metadata, three of these clusters were characterized as highly eosinophilic, three as highly neutrophilic, and two as highly atopic with relatively low levels granulocytes. Remarkably, high levels IL-13 and periostin, but low total serum IgE levels were found in clusters with the highest eosinophilia, whereas the most atopic proteotypes showed high levels of total IgE, but moderate levels of IL-13 and periostin, suggesting different underlying mechanisms for some of type 2 asthma variants. For each of these three phenotypes, candidate protein biomarkers were identified and matched transcriptomic data pointed to differentially activated underlying mechanisms .
Only very limited proteomic data are available in asthmatic children. A recent study identified IL-8 and IL-10 in the saliva of asthmatic children as biomarkers of bronchial inflammation and obstruction . Another study showed that children with severe asthma have decreased lipoxin A4 (LXA4) concentrations in sputum in comparison with children with intermittent asthma. LXA4 concentrations negatively correlated with leukotriene B4 concentrations and with exacerbation numbers in children with severe asthma. The same study demonstrated in vitro a crosstalk between LXA4 and glucocorticoid receptor at the cytosolic level in peripheral blood granulocytes isolated from children with asthma, which may point towards the mechanism behind the reduction in the ability of ICS to impair control of airway inflammation in children with severe asthma . Moreover, a statistically significant negative correlation between FEV1/FVC and sputum neutrophil gelatinase-associated lipocalin and matrix metalloproteinase-9 in obese asthmatic children was demonstrated .
TOWARDS INTEGRATION OF OMICS DATA
The next step should be the integration of omics data with clinical and environmental data. However, a big challenge is the size of the data together with differences in nomenclature among data types, which will require novel analytical procedures and the collaboration with data scientists. Commonly used approaches are currently limited by the three I's – integration, interpretation, and insights. Post integration, these large datasets aim to yield views of cellular systems at high resolution for transformative insights into processes, events, and diseases through various computational and informatics frameworks [37,38]. Single omic analyses may provide some insight into the basis of lung function in children with asthma, but the underlying biologic pathways are still poorly understood. Kelly et al. expanded the single omic findings by integrating the previously correlated gene-metabolite modules collected from 1165 asthmatic children. Weighted gene co-expression network analysis clustered 25 060 gene probes and 8185 metabolite features into eight gene modules and eight metabolite modules, where four and six, respectively, were associated with lung function. While gene modules were enriched for immune, mitotic, and metabolic processes and asthma-associated microRNA targets, the metabolite modules were enriched for lipid and amino acid metabolism . Strategies to integrate risk factors from multiple distinct data sets have been hampered by the issue of missing data and lack of methods to deal with that. Krautenbacher et al. collected and analyzed different types of data (questionnaires, diagnostic, genotype, microarray, RT-qPCR, flow cytometry, and cytokine data) from healthy children, mild-to-moderate allergic asthmatics, and nonallergic asthmatics from 260 German children aged 4–14 years. These data were used for building a novel multilevel prediction approach for childhood asthma phenotyping and outcome, which could deal with a missing data structure. Remarkably, this study identified 4PKN2 (protein kinaseN2), PTK2 (protein tyrosine kinase 2), and ALPP (alkaline phosphatase, placental) as the most important variables for classifying childhood asthma phenotype .
UNMET NEEDS AND FUTURE PERSPECTIVES
Despite the large amount of data the new omics era has provided on type 2 high and type 2 low driven patterns of airway inflammation, the only biomarkers that are currently recommended in the GINA guidelines, as well as a recent ERS/ATS Task Force on the management of severe asthma to guide treatment with biologicals: are total and/or specific serum IgE, blood eosinophils, and FeNO [3,41]. Since these recommendations are mainly based on adult studies, it remains unclear how recommendations can be translated to the younger population, since studies in children are scarce and results from adult studies should only be extrapolated with caution. The clinical expression of severe asthma in children differs from adults, but also normal ranges of biomarkers might be dependent on age, that is blood eosinophilia is observed relatively frequently within the pediatric population .
To apply the latest omics technology to guide treatment in asthmatic children (Fig. 1), there is a need for biomarker discovery studies on biologics in children, validation studies, biomarker-guided studies, and implementation studies. So far, most published biomarker studies are in the phase of biomarker discovery for asthma prediction, asthma severity, or corticosteroid response. These studies often report associations. Validation studies should include functional validation of biomarkers using ex-vivo models, development of targeted panels of selected omics markers preferably incorporated in a simple clinical test or treatment algorithm, and clinical validation in independent cohorts. Subsequently, biomarker-guided trials need to be performed to assess whether such a clinical test or algorithm outperforms current clinical practice. Lastly, implementation studies and economic evaluations are needed to assess the clinical utility.
In addition, there is increasing awareness of the role of the airways and gut/airways microbiota in the onset and course of asthma. As microbiome of the gut was shown to participate in the shaping of the immune responses at distal sites, i.e. in the lung, it has become an interesting subject of research with respect to finding associations between microbiota composition and asthma development or severity.
To date, several cross-sectional and longitudinal microbiome studies have been conducted in asthmatics with different phenotypes, which have revealed many associations between bacterial composition and asthma . Similarly, microbial networks of co-occurrence of bacterial genera revealed different bacterial associations across asthma phenotypes . An increasing number of studies reporting on microbiome association with asthma leaves us with a promising tool that in the near future may allow implementing a microbiomic phenotype as a tool for the prediction of disease progression of responses to medications.
To move omics advances towards implementation, we recently established the PERsonalized Medicine Approach for asthma and allergy Biological SeLEction (PERMEABLE) consortium. This multidisciplinary European consortium aims to establish consensus on clinical selection criteria for young biologicals users and combine preclinical studies on treatment response with multiomics biomarker studies of young patients starting with a biological. To incorporate precision medicine to the pediatric asthma clinic, we need to identify and validate predictors of nonresponse to corticosteroids, as well as predictors for response to novel targeted treatments. These predictors should be accurate and preferably as noninvasive as possible. Collaboration and bringing experts from different fields together (such as clinicians, pharmacologists, immunologists, and data scientists) is inevitable to pave the way for more precise, personalized, and effective management of childhood asthma.
Omics studies have largely expanded our view on asthma heterogeneity, helped understand cellular processes underlying asthma, and brought us closer towards identifying (bio)markers that will allow the prediction of treatment responsiveness and disease progression. A next promising step could be the integration of multi omics data together with clinical data, to feed into an individual risk prediction model. However, this requires large-scale collaboration in a multidisciplinary setting.
The PERMEABLE consortium wishes to acknowledge the help and expertise of the following individuals without whom, this work would not have been possible: Heike Buntrock-Döpke and Elisangela Santos-Valente (University Children's Hospital Regensburg (KUNO), Regensburg, Germany); Jakob Nigel and Gabriel Zink (Byteschmiede UG, Parsberg, Germany); Boris Gole (Molecular Biology and Genomics, Faculty for Chemistry and Chemical Engineering, University of Maribor, Maribor, Slovenia); Niels Rutjes (Pediatric Pulmonology, Amsterdam UMC, University of Amsterdam, Amsterdam, the Netherlands), Paul Brinkman, Simone Hashimoto (Respiratory Medicine, Amsterdam UMC, University of Amsterdam, the Netherlands); Jon Konradsen (Department of Women's anc Children's Health, Karolinska Institutet, Stockholm, Sweden), Caroline Nilsson (Department of Clinical Sciences and Education, Södersjukhuset, Karolinska Institutet, Stockholm, Sweden).
Financial support and sponsorship
The PERMEABLE consortium is supported by ZonMW (project number: 456008004), the Swedish Research Council (project number 2018-05619), the Ministry of Education, Science and Sport of the Republic of Slovenia (contract number C3330-19-252012), the German Ministry of Education and Research (BMBF) (project number FKZ01KU1909A), under the frame of the ERA PerMed JTC 2018 Call.
Conflicts of interest
There are no conflicts of interest.
REFERENCES AND RECOMMENDED READING
Papers of particular interest, published within the annual period of review, have been highlighted as:
1. Agache I, Cojanu C, Rogozea L. Agache I, Hellings P. Chapter 8 – Endotype-driven approach for asthma. Implementing precision medicine
in best practices of chronic airway diseases. Karnataka, India: Academic Press; 2019. 45–49.
2▪. Tyler SR, Bunyavanich S. Leveraging -omics
for asthma endotyping. J Allergy Clin Immunol 2019; 144:13–23.
3. Holguin F, Cardet JC, Chung KF, et al. Management of severe asthma: a European Respiratory Society/American Thoracic Society Guideline. Eur Respir J 2020; 55:
4. Moffatt MF, Kabesch M, Liang L, et al. Genetic variants regulating ORMDL3 expression contribute to the risk of childhood asthma
. Nature 2007; 448:470–473.
5. Stein MM, Thompson EE, Schoettler N, et al. A decade of research on the 17q12-21 asthma locus: piecing together the puzzle. J Allergy Clin Immunol 2018; 142:749–764.e3.
6. Demenais F, Margaritte-Jeannin P, Barnes KC, et al. Multiancestry association study identifies new asthma risk loci that colocalize with immune-cell enhancer marks. Nat Genet 2018; 50:42–53.
7. Dijk FN, Folkersma C, Gruzieva O, et al. Genetic risk scores do not improve asthma prediction in childhood. J Allergy Clin Immunol 2019; 144:857–860.e7.
8. Sugier PE, Sarnowski C, Granell R, et al. Genome-wide interaction study of early-life smoking exposure on time-to-asthma onset in childhood. Clin Exp Allergy 2019; 49:1342–1351.
9. Gref A, Merid SK, Gruzieva O, et al. Genome-wide interaction analysis of air pollution exposure and childhood asthma
with functional follow-up. Am J Respir Crit Care Med 2017; 195:1373–1383.
10▪▪. Vijverberg SJH, Farzan N, Slob EMA, et al. Treatment response heterogeneity in asthma: the role of genetic variation. Exp Rev Respir Med 2018; 12:55–65.
11. Farzan N, Vijverberg SJ, Hernandez-Pacheco N, et al. 17q21 variant increases the risk of exacerbations in asthmatic children despite inhaled corticosteroids use. Allergy 2018; 73:2083–2088.
12. Berce V, Kozmus CE, Potocnik U. Association among ORMDL3 gene expression, 17q21 polymorphism and response to treatment with inhaled corticosteroids in children with asthma. Pharmacogenomics J 2013; 13:523–529.
13. Turner S, Francis B, Vijverberg S, et al. Childhood asthma
exacerbations and the Arg16 beta2-receptor polymorphism: a meta-analysis stratified by treatment. J Allergy Clin Immunol 2016; 138:107–113.e5.
14. Vijverberg SJ, Pijnenburg MW, Hövels AM, et al. The need for precision medicine
clinical trials in childhood asthma
: rationale and design of the PUFFIN trial. Pharmacogenomics 2017; 18:393–401.
15. Forno E, Celedon JC. Epigenomics and transcriptomics in the prediction and diagnosis of childhood asthma
: are we there yet? Front Pediatr 2019; 7:115.
16▪. Xu CJ, Soderhall C, Bustamante M, et al. DNA methylation in childhood asthma
: an epigenome-wide meta-analysis. Lancet Respir Med 2018; 6:379–388.
17▪▪. Reese SE, Xu CJ, den Dekker HT, et al. Epigenome-wide meta-analysis of DNA methylation and childhood asthma
. J Allergy Clin Immunol 2019; 143:2062–2074.
18. Cardenas A, Sordillo JE, Rifas-Shiman SL, et al. The nasal methylome as a biomarker of asthma and airway inflammation in children. Nat Commun 2019; 10:3095.
19. Wang AL, Qiu W, DeMeo DL, et al. DNA methylation is associated with improvement in lung function on inhaled corticosteroids in pediatric asthmatics. Pharmacogenet Genomics 2019; 29:65–68.
20. Kere M, Gruzieva O, Ullemar V, et al. Effects of inhaled corticosteroids on DNA methylation in peripheral blood cells in children with asthma. Allergy 2019; [Epub ahead of print].
21▪. Altman MC, Gill MA, Whalen E, et al. Transcriptome networks identify mechanisms of viral and nonviral asthma exacerbations in children. Nat Immunol 2019; 20:637–651.
22. Qiu W, Guo F, Glass K, et al. Differential connectivity of gene regulatory networks distinguishes corticosteroid response in asthma. J Allergy Clin Immunol 2018; 141:1250–1258.
23. Hekking PP, Loza MJ, Pavlidis S, et al. Pathway discovery using transcriptomic profiles in adult-onset severe asthma. J Allergy Clin Immunol 2018; 141:1280–1290.
24. Jiang Y, Gruzieva O, Wang T, et al. Transcriptomics of atopy and atopic asthma in white blood cells from children and adolescents. Eur Respir J 2019; 53:
25▪. Melen E, Guerra S, Hallberg J, et al. Linking COPD epidemiology with pediatric asthma care: implications for the patient and the physician. Pediatr Allergy Immunol 2019; 30:589–597.
26. Tibbitt CA, Stark JM, Martens L, et al. Single-cell RNA sequencing of the T helper cell response to house dust mites defines a distinct gene expression signature in airway Th2 cells. Immunity 2019; 51:169–184.e5.
27▪▪. Neerincx AH, Vijverberg SJH, Bos LDJ, et al. Breathomics from exhaled volatile organic compounds in pediatric asthma. Pediatr Pulmonol 2017; 52:1616–1627.
28. de Vries R, Dagelet YWF, Spoor P, et al. Clinical and inflammatory phenotyping by breathomics in chronic airway diseases irrespective of the diagnostic label. Eur Respir J 2018; 51:
29▪. Brinkman P, van de Pol MA, Gerritsen MG, et al. Exhaled breath profiles in the monitoring of loss of control and clinical recovery in asthma. Clin Exp Allergy 2017; 47:1159–1169.
30. van Bragt JMH, de Vries R, Sterk PJ, Maitland-van der Zee AH. Exhaled breath analysis for prediction of responders to mepolizumab in patients with severe asthma. Am J Respir Crit Care Med 2019; 199:A2673.
31. Robroeks CM, van Berkel JJ, Jobsis Q, et al. Exhaled volatile organic compounds predict exacerbations of childhood asthma
in a 1-year prospective study. Eur Respir J 2013; 42:98–106.
32. van Vliet D, Smolinska A, Jobsis Q, et al. Can exhaled volatile organic compounds predict asthma exacerbations in children? J Breath Res 2017; 11:016016.
33. Schofield JPR, Burg D, Nicholas B, et al. Stratification of asthma phenotypes by airway proteomic signatures. J Allergy Clin Immunol 2019; 144:70–82.
34. Zamora-Mendoza BN, Espinosa-Tanguma R, Ramirez-Elias MG, et al. Surface-enhanced raman spectroscopy: a non invasive alternative procedure for early detection in childhood asthma
biomarkers in saliva. Photodiagnosis Photodyn Ther 2019; 27:85–91.
35. Gagliardo R, Gras D, La Grutta S, et al. Airway lipoxin A4/formyl peptide receptor 2-lipoxin receptor levels in pediatric patients with severe asthma. J Allergy Clin Immunol 2016; 137:1796–1806.
36. Nacaroglu HT, Gayret OB, Erol M, et al. Biomarkers of airway and systemic inflammation in obese asthmatic paediatric patients. Allergol Immunopathol 2017; 45:534–540.
37. Misra BB, Langefeld CD, Olivier M, Cox LA. Integrated omics
: tools, advances, and future approaches. J Mol Endocrinology 2018; 62:21–45.
38. Ivanova O, Richards LB, Vijverberg SJ, et al. What did we learn from multiple omics
studies in asthma? Allergy 2019; 74:2129–2145.
39. Kelly RS, Chawes BL, Blighe K, et al. An integrative transcriptomic and metabolomic study of lung function in children with asthma. Chest 2018; 154:335–348.
40. Krautenbacher N, Flach N, Bock A, et al. A strategy for high-dimensional multivariable analysis classifies childhood asthma
phenotypes from genetic, immunological, and environmental factors. Allergy 2019; 74:1364–1373.
41. Global Initiative for Asthma (GINA). Pocket guide ‘diagnosis and management of difficult-totreat and severe asthma in adolescent and adult patients’ website: www.ginasthma.org
. [Accessed 30 November 2019].
42. Schwartz JT, Fulkerson PC. An approach to the evaluation of persistent hypereosinophilia in pediatric patients. Front Immunol 2018; 9:1944.
43. Abdel-Aziz MI, Vijverberg SJH, Neerincx AH, et al. The crosstalk between microbiome and asthma: exploring associations and challenges. Clin Exp Allergy 2019; 49:1067–1086.
44. Perez-Losada M, Authelet KJ, Hoptay CE, et al. Pediatric asthma comprises different phenotypic clusters with unique nasal microbiotas. Microbiome 2018; 6:179.