The Genomic Revolution in Pulmonary Medicine: A Comprehensive Narrative Review of Genomic and Multi-Omic Technologies in Respiratory Conditions

Abstract

Chronic respiratory diseases, including chronic obstructive pulmonary disease (COPD), asthma, and interstitial lung diseases (ILDs), represent a major global health burden. Their significant clinical and biological heterogeneity complicates diagnosis and limits the efficacy of traditional, one-size-fits-all management approaches. The advent of high-throughput genomic and multi-omic technologies has initiated a paradigm shift from syndromic classification to molecular-based endotyping. A narrative review of the literature was performed, synthesising foundational and recent research in the genomics, epigenomics, and multi-omics of chronic respiratory diseases. Key studies were selected based on their relevance to genetic architecture, biomarker development, and translational applications in precision medicine. We discuss the complex genetic architecture of pulmonary conditions, highlighting the contribution of both rare, high-penetrance variants, such as SERPINA1, CFTR, and BMPR2, and polygenic risk from many common variants, such as HHIP, FAM13A, and IL33. We provide detailed analyses of polygenic risk scores (PRSs) for COPD and asthma, including their construction, validation across ancestries, and predictive performance. We detail how integrative multi-omic approaches, including transcriptomics, proteomics, and metabolomics, are successfully defining molecular endotypes, such as Type 2-high asthma, which, in turn, inform the use of targeted biologic therapies. Finally, we review the development of molecular diagnostics, including metagenomic sequencing of infections and liquid biopsies for lung cancer and the development of prognostic biomarkers. The genomic revolution is transforming pulmonary medicine through the discovery of novel disease pathways, precise molecular classification, and the recognition of new therapeutic targets. Despite major challenges in functional interpretation, data integration, and clinical–translational equity, these technologies hold the key to a new era of personalised respiratory health and precision medicine.

1. Introduction

Chronic respiratory diseases, such as chronic obstructive pulmonary disease (COPD), asthma, and interstitial lung diseases (ILDs), are among the most prevalent and burdensome non-communicable diseases globally [1]. These conditions contribute significantly to global morbidity and mortality, with asthma and COPD alone affecting over 500 million individuals and causing millions of deaths each year. Recent analyses have demonstrated worrying trends in chronic respiratory disease burden across the world [2,3]. Although symptomatic management has advanced, these diseases remain incurable, and their clinical progression is often unpredictable due to underlying biological complexity and heterogeneity [4].

Traditionally, diagnosis and management of respiratory diseases have relied on syndromic frameworks that include clinical symptoms, pulmonary function testing, and imaging. These approaches are useful but often group etiologically distinct disorders into similar categories, masking underlying disease mechanisms [5]. For example, asthma and COPD each consist of multiple clinical phenotypes and molecular endotypes, which are distinguished by overlapping but distinct inflammatory pathways, immune responses, and structural changes to the airways. This heterogeneity explains the wide variation in disease trajectories and therapeutic responses among patients with the same clinical diagnosis [5,6].

The recognition of this mechanistic heterogeneity has led to a paradigm shift towards molecular and genomic classification in pulmonary medicine. Genome-wide association studies (GWASs) and whole-genome sequencing (WGS) have identified hundreds of loci associated with lung function, COPD, asthma, and related traits [7]. Studies have demonstrated the highly polygenic nature of chronic respiratory diseases, in which many variants of small individual effects collectively influence disease susceptibility, progression, and response to environmental exposures such as smoking and air pollution, as well as treatment options [8]. For instance, the COPDGene study and multi-ancestry GWAS have identified over 20 loci associated with COPD and more than 1000 independent association signals for lung function, implicating hundreds of genes and dozens of biological pathways, including those related to immune regulation, tissue remodelling, and nicotinic acetylcholine receptor signalling [9].

Genetic risk factors show considerable overlap among respiratory diseases, indicating shared biological mechanisms and gene–environment interactions [10]. Variants in genes that regulate the immune response, epithelial integrity, and airway structure are implicated across both asthma and COPD; these developmental pathways of the lung contribute to disease susceptibility [11]. These insights can lead to improved risk assessment, early diagnosis, and stratification of treatment options.

Beyond DNA sequence variations, epigenetic mechanisms including DNA methylation, histone modifications, and non-coding RNAs act as dynamic mediators between the genetic background and environmental exposures [12]. Aberrant epigenetic signatures have been identified in COPD, asthma, and pulmonary arterial hypertension (PAH), pointing to reversible changes in gene expression that drive pathogenesis and heterogeneity [13]. Epigenetic mechanisms are influenced by smoking, pollution, and early-life exposures and can be detected in blood, sputum, or airway tissues, thus serving as potential non-invasive biomarkers. The integration of epigenomic data with genetic and transcriptomic information, using network and systems biology approaches, is providing new insights into the multistage transition from health to disease and the identification of pre-disease states amenable to intervention [13,14].

Newer technologies including transcriptomics, proteomics, metabolomics, and microbiomics further expand this molecular view. They allow for the comprehensive profiling of inflammatory pathways (e.g., type 2, type 1, and type 17 responses), immune cell signatures, and metabolic changes driving disease heterogeneity [15,16]. For example, comprehensive profiling has highlighted the predominance of type 2 inflammation in asthma and COPD phenotypes while also identifying steroid-refractory phenotypes and novel markers like interleukin-6 (IL-6). MicroRNAs (miRNAs), as key post-transcriptional regulators, have emerged as promising diagnostic and prognostic biomarkers, with deregulation observed in multiple biofluids and tissues from patients with asthma and COPD [15,16].

Integration of this new genetic data now enables the identification of new molecular endotypes, risk stratification tools, and novel therapeutic targets. Polygenic risk scores (PRSs) and molecular biomarkers have the potential to predict disease onset, progression, and treatment response. These advances form the basis for the shift from a ‘one-size-fits-all’ model to ‘precision medicine’ in pulmonary diseases. In this review, we provide a comprehensive and detailed analysis of the key genomic and DNA-based technologies that are shaping pulmonary medicine, focusing on their applications, challenges, and translational prospects. An overview of these technologies and their applications is presented in Figure 1.

2. Genetic Architecture of Respiratory Diseases

2.1. Heritability and the Polygenic Basis of Respiratory Disease

Respiratory diseases are highly heritable, with estimates for lung function and COPD ranging from 38% to 50% [7,17]. However, the contribution of individual variants accounts for only a small portion of overall risk [18]. For instance, a total of 279 variants identified via a large-scale GWAS explain approximately 13% of lung function heritability, with the contribution of many small-effect variants and rare high-impact alleles also implicated [7]. This ‘missing heritability’ is attributed to either common variants contributing small effects that miss GWAS significance thresholds or rare variants contributing to disease risk in a small portion of the population [18]. Understanding this polygenic architecture is essential for developing risk prediction tools and for appreciating why single-gene approaches are insufficient for most respiratory diseases. The shared and disease-specific genetic architecture across respiratory diseases is illustrated in Figure 2.

2.2. Polygenic Risk Scores for COPD: Construction, Validation, and Clinical Implications

The development of the PRS for COPD represents one of the most clinically promising applications of respiratory genomics. The PRS aggregates the effects of millions of genetic variants, each contributing a small individual effect, into a single quantitative measure of inherited disease risk. The evolution of the COPD PRS illustrates how increasing GWAS sample sizes and methodological refinements have progressively improved predictive performance.

2.2.1. Construction and Methodological Evolution

Early genetic risk scores for COPD were constructed from a limited number of genome-wide significant variants. An initial GWAS of approximately 12,000 individuals yielded an unweighted 30-variant genetic risk score with modest predictive power (area under the curve [AUC]: 0.58) [19]. A subsequent GWAS including nearly 49,000 UK Biobank participants produced a 95-variant risk score with a 3.7-fold greater risk of COPD when comparing the highest and lowest deciles [7]. The largest multi-ancestry GWAS, including more than 400,000 individuals, identified 279 variants and generated a weighted genetic risk score with 4.73-fold increased odds of COPD in the highest versus lowest decile [7]. The most comprehensive PRS to date, developed by Moll et al. using a Bayesian framework (LDpred) applied to GWAS summary statistics from over 400,000 participants, incorporated more than 2 million variants weighted by their full GWAS effect estimates rather than limiting inclusion to genome-wide significant variants [19]. This combined FEV1 and FEV1/FVC PRS demonstrated 8.0-fold increased odds of COPD in the highest versus lowest decile in European populations, with a meta-analysed AUC of 0.68 [19]. It should be noted that odds ratios, particularly when comparing extreme deciles, can assume large numerical values that may overstate the absolute risk difference. These ORs reflect relative genetic enrichment rather than absolute clinical risk, and their interpretation should account for the baseline prevalence of COPD in the population.

2.2.2. Clinical Associations Beyond Diagnosis

The COPD PRS is not merely a diagnostic tool; it captures meaningful biological heterogeneity. The PRS correlates with emphysema burden and airway disease on quantitative computed tomography (CT) imaging and with patterns of reduced lung growth in children with asthma followed through to adulthood, raising the possibility that the PRS may reflect risk of specific disease features rather than just overall COPD risk [1,4]. Zhang et al. demonstrated that the PRS is associated with earlier age of COPD diagnosis, with age-dependent effects that are larger at younger ages, and that adding the PRS to known early-life risk factors (e.g., childhood asthma) improved prediction of COPD before age 50 (AUC improvement from 0.69 to 0.74 in non-Hispanic Whites) [5]. A recent study combining the PRS with clinical factors such as age, sex, and smoking achieved an AUC of approximately 0.80 for identifying undiagnosed COPD.

2.2.3. Gene–Smoking Interactions

Zhang et al. investigated the interaction between the PRS and smoking in 439,255 UK Biobank participants and found that individuals with high genetic risk who were current smokers had a hazard ratio of 11.62 for incident COPD compared with low-risk never-smokers [7]. The population-attributable risk of smoking increased from 42.7% in the low-genetic-risk group to 61.1% in the high-genetic-risk group, demonstrating that genetic susceptibility amplifies the harmful effects of smoking [7].

2.2.4. Limitations and Equity Considerations

Most COPD PRSs have been derived and validated predominantly in European-ancestry cohorts. Performance in non-European populations is attenuated; for example, the PRS showed weaker associations in African American participants (OR 1.23 vs. 1.55 per SD in non-Hispanic Whites for COPD before age 50) [5]. Expanding multi-ancestry GWASs and developing population-specific PRSs are critical priorities for ensuring equitable clinical application. In our view, the COPD PRS is approaching clinical utility as a complementary risk stratification tool, particularly for identifying high-risk individuals who might benefit from early screening, smoking cessation interventions, and longitudinal monitoring. However, it does not yet replace clinical assessment and should be integrated with, rather than substituted for, traditional risk factors.

2.3. Genetic Risk Scores for Asthma: Current State and Future Potential

Asthma exhibits a strong polygenic component, with GWAS signals mapping to immune-regulatory and epithelial genes such as IL33 and TSLP, which are now recognised to be central to T helper 2 (Th2) inflammation [20]. The development of the asthma PRS has followed a trajectory similar to that of the COPD PRS but with unique challenges related to disease heterogeneity and age of onset.

Sordillo et al. developed an asthma PRS using both a standard approach (genome-wide significant variants) and lasso sum regression approach (allowing all variants to contribute) in the racially diverse Kaiser Permanente GERA cohort (68,638 non-Hispanic Whites, 5874 Hispanics, 6870 Asians, and 2760 Blacks) [10]. The lasso sum regression PRS showed stronger associations across all groups: ORs per SD of 1.20 in non-Hispanic Whites, 1.17 in Hispanics, 1.18 in Asians, and 1.10 in Blacks [10]. Namjou et al. developed a multi-ancestral PRS using a Bayesian regression framework and demonstrated significant discrimination across paediatric subcohorts of European (AUC: 0.60–0.66), African (AUC: 0.61–0.66), admixed American (AUC: 0.64–0.70), and East Asian (AUC: 0.73) ancestry, with participants in the top 5% PRSs having 2.80- to 5.82-fold increased odds of asthma compared with the bottom 5% [9].

A critical insight from recent PRS studies is that genetic risk scores perform better for childhood-onset asthma than adult-onset asthma, consistent with the stronger genetic basis of early-onset disease. Koppelman et al. reported that the ORs per SD for childhood-onset asthma versus adult-onset asthma were 1.75 versus 1.24 in European-descent populations, with AUC values exceeding 0.70 only for childhood-onset asthma among European, Hispanic/Latino, and East Asian populations [12]. This finding has important implications: the current asthma PRS may be most useful for identifying children at high genetic risk who could benefit from early intervention or prevention strategies.

Moll et al. demonstrated that PRSs for asthma and spirometry have differential associations with asthma, COPD, and asthma–COPD overlap (ACO) [11]. Both PRSs predicted ACO, and the spirometry PRS was associated with asthma exacerbations across multiple racial/ethnic groups, suggesting that genetic risk scores may help disentangle the complex relationship between these overlapping conditions [11]. Pariès et al. showed that in a multivariate analysis, PRSs provided decision-making information similar to that of airway hyperresponsiveness testing, with the top 10% and bottom 5% of PRSs being the most informative thresholds for defining high and low genetic risk [14].

We believe that the asthma PRS holds particular promise for paediatric risk stratification and for identifying individuals at risk of severe or exacerbation-prone disease. However, the heterogeneity of asthma spanning allergic, eosinophilic, and non-eosinophilic endotypes means that a single PRS may not capture all dimensions of disease risk and an endotype-specific PRS may ultimately prove more clinically useful.

2.4. Key Genetic Variants and Loci in COPD

The genetics of COPD susceptibility comprises both rare and common variants. The best-characterised rare variant is the mutation in SERPINA1 causing alpha-1 antitrypsin deficiency (AATD), which is a major determinant of early-onset panacinar emphysema [21]. Hundreds of common loci have been mapped near HHIP, FAM13A, CHRNA3/5, IREB2, RIN3, CYP2A6, DSP, MMP12, and many others [19]. The most recent multi-ancestry GWAS identified over 1000 independent signals implicating more than 500 genes and approximately 30 pathways enriched in lung, smooth muscle, and immune cells [7]. Many of these loci overlap with those for pulmonary fibrosis, suggesting shared genetic mechanisms across obstructive and restrictive lung diseases [19].

2.5. Interstitial Lung Diseases

The genetic architecture of ILDs, especially idiopathic pulmonary fibrosis (IPF), comprises common and rare variants. The MUC5B promoter variant (rs35705950) is the strongest common genetic risk factor for IPF, with an odds ratio of approximately 5–6 per allele [20]. Shared loci with COPD and asthma include FAM13A and DSP, though these loci sometimes confer opposite effects on risk across diseases [19,20]. Rare variants in telomere-related genes (TERT, TERC, RTEL1, PARN) and surfactant-pathway genes (SFTPC, SFTPA2) are found in familial and sporadic pulmonary fibrosis, with telomere-related variants present in up to 25% of familial IPF cases [22,23]. These findings have direct clinical implications: short telomere length is associated with worse outcomes after lung transplantation and may influence tolerance of immunosuppressive therapy; fibrotic ILDs also carry significant pulmonary vascular consequences [24].

2.6. Pulmonary Arterial Hypertension

PAH is a serious vascular disorder with a strong genetic basis. Heterozygous BMPR2 mutations are present in approximately 80% of familial cases and about 20% of sporadic cases [25]. An individual participant data meta-analysis demonstrated that BMPR2 mutation carriers present at a younger age and with more severe haemodynamic compromise, though survival after diagnosis does not differ significantly from non-carriers [25]. Other pathogenic variants occur in ACVRL1, ENG, SMAD family members, CAV1, KCNK3, TBX4, SOX17, ATP13A3, and AQP1, which converge on TGF-β/BMP signalling [26]. Biallelic EIF2AK4 mutations cause pulmonary veno-occlusive disease [27]. These findings have direct implications for genetic counselling, cascade screening of at-risk family members, and targeted-therapy development.

2.7. Cystic Fibrosis: The Paradigm of Genotype-Directed Therapy

Cystic fibrosis (CF) results from pathogenic variants in the CFTR gene encoding a chloride channel crucial for epithelial fluid transport. More than 2000 CFTR variants have been described, with F508del being by far the most common in North America, present in approximately 70% of CF alleles [27]. Ethnic variation in mutation spectra impacts diagnosis and eligibility for therapy [28].

CF has become the paradigm for genotype-directed therapy in pulmonary medicine. CFTR modulators are specifically matched to mutation classes: ivacaftor (a potentiator) is effective for gating mutations such as G551D, while the triple combination elexacaftor–tezacaftor–ivacaftor (ETI) targets the folding and gating defects of F508del [29,30]. In the pivotal phase 3 trial, ETI improved ppFEV1 by 10.0 percentage points, decreased sweat chloride by 45.1 mmol/L, and improved CFQ-R respiratory domain scores by 17.4 points compared with tezacaftor–ivacaftor alone in F508del-homozygous patients [31]. In patients heterozygous for F508del and a minimal function mutation previously ineligible for modulator therapy, ETI reduced annualised pulmonary exacerbation rates by 63% (rate ratio: 0.37) and improved ppFEV1 by 14.3 percentage points [32]. The concept of ‘theratyping’ using in vitro cell-based assays to identify responsive variants has expanded ETI eligibility to approximately 90% of people with CF, including those with rare variants [32,33]. This approach, in which a patient’s specific genotype directly determines therapeutic eligibility and expected response, represents the most mature example of precision medicine in pulmonary disease.

2.8. Sleep Apnoea

Obstructive sleep apnoea (OSA) results from both environmental and genetic risk factors [34]. GWASs and transcriptome-wide analyses point to genes such as HTR1F, GPD2, and L3MBTL2, which implicate serotonin signalling, metabolism, and DNA repair pathways [19,35]. This mechanistic diversity points to neuronal and metabolic regulation in OSA and suggests potential pharmacogenomic approaches to treatment.

2.9. Lung Cancer

Risk of lung cancer emanates from both environmental exposures, particularly smoking, and inherited predisposition. Family history alone is independently associated with heightened risk [36]. GWASs have identified variants at chromosomal regions 5p15.33 and 3q28 and rare high-penetrance mutations in EGFR, YAP1, and DNA repair genes, including BRCA2, CHEK2, ATM, and BRCA1 [37,38,39]. Clonal haematopoiesis and somatic driver mutations add further heterogeneity and directly inform strategies for therapeutic targeting [37]. Translating these genomic insights into clinical benefit, however, remains uneven [40].

2.10. Integrative Insights Across Respiratory Diseases

Genetic studies in pneumonia, hypersensitivity pneumonitis, and sarcoidosis implicate MHC class II HLA alleles [38,39]. Telomere-related and surfactant-pathway genes are involved in fibrotic ILDs [22,23]. Shared architecture among IPF, COPD, and sleep apnoea underlines inflammation-related mucin gene clusters and immune pathways [19]. Across respiratory diseases, GWASs and sequencing together implicate hundreds of loci enriched in immune, epithelial, and developmental pathways [7,41]. Rare variants, ethnic diversity, and gene–environment interactions represent key research frontiers [42].

Table 1. Summary of genetic architecture across key respiratory diseases.

Disease	Heritability	Key Rare Variants	Key Common Loci	PRS Performance	Clinical Implications
COPD	38–50%	SERPINA1 (AATD)	HHIP, FAM13A, CHRNA3/5, IREB2, DSP, MMP12	AUC: 0.68–0.80; OR: up to 8.0× (top vs. bottom decile)	Early screening; smoking cessation; PRS + clinical model AUC ≈ 0.80
Asthma	60–80% (childhood)	Rare TSLP, IL33 variants	IL33, TSLP, ORMDL3, HLA-DQ, IL13, GSDMB	AUC: 0.60–0.73; OR: 1.10–1.75 per SD	Paediatric risk stratification; biologic therapy selection
IPF	~30–40%	TERT, TERC, RTEL1, PARN, SFTPC, SFTPA2	MUC5B (OR 5–6), FAM13A, DSP, TOLLIP	No formal PRS; MUC5B OR: 5–6 per allele	Telomere testing; transplant risk; antifibrotic eligibility
PAH	~70–80% (familial)	BMPR2 (~80% familial, ~20% sporadic), ACVRL1, ENG	TBX4, SOX17, ATP13A3	BMPR2 mutation: earlier onset, severe haemodynamics	Genetic counselling; cascade family screening
CF	Near 100% (monogenic)	CFTR (>2000 variants; F508del ~70% of alleles)	Modifier genes: MSRA, SLC26A9	Genotype determines therapy eligibility	Genotype-directed CFTR modulators (ETI); theratyping
OSA	~40–60%	None established	HTR1F, GPD2, L3MBTL2	Early stage; no clinical PRS	Pharmacogenomic potential; serotonin pathway targets
Lung Cancer	~15–20%	BRCA2, CHEK2, ATM, BRCA1; EGFR, YAP1	5p15.33 (TERT), 3q28 (TP63), 15q25	Family history: OR: ~2×	Liquid biopsy; somatic driver profiling; targeted therapy

AATD—alpha-1 antitrypsin deficiency; CF—cystic fibrosis; COPD—chronic obstructive pulmonary disease; ETI—elexacaftor–tezacaftor–ivacaftor; IPF—idiopathic pulmonary fibrosis; OR—odds ratio; OSA—obstructive sleep apnoea; PAH—pulmonary arterial hypertension; PRS—polygenic risk score; SD—standard deviation; AUC—area under the curve.

summarizes the genetic architecture across key respiratory diseases.

3. Epigenetic and Gene–Environment Interactions

Epigenetic and gene–environment interactions are key to the pathogenesis, progression, and heterogeneity of pulmonary diseases [43]. Epigenetic regulation through DNA methylation, histone modification, and non-coding RNAs allows for the modulation of gene expression without changes in DNA sequence [44]. These mechanisms respond dynamically to exposures such as tobacco smoke, air pollution, infection, and diet [45,46]. The major epigenetic mechanisms linking environmental exposures to respiratory disease are depicted in Figure 3.

3.1. DNA Methylation

DNA methylation is the most widely studied epigenetic modification in lung disease [43]. Aberrant methylation patterns have been found in COPD, asthma, IPF, PAH, CF, and lung cancer, impacting inflammation, apoptosis, immunity, and tissue remodelling [1,47]. Epigenome-wide association studies (EWASs) have identified reproducible differentially methylated positions (DMPs) in both blood and respiratory tissues [1]. A systematic review by Casas-Recasens et al. identified 51 replicated DMPs in blood related to lung function and 12 related to COPD, while 42 DMPs were replicated in respiratory samples [46]. Importantly, only 2.6% of total genes were shared between blood and respiratory samples, suggesting that blood can recapitulate some but not all changes in respiratory tissues [46]. Eriksson Ström et al. performed the first EWAS on bronchoalveolar lavage (BAL) cells and found 1155 Bonferroni-significant DMPs associated with COPD, with 39% colocalising with COPD-associated single-nucleotide polymorphisms (SNPs), suggesting joint genetic and epigenetic pathways [47]. Differentially methylated sites in COPD are enriched adjacent to COPD-associated GWAS loci, suggesting that genetic regulation of gene expression may occur through DNA methylation [1]. Such changes, detectable in peripheral blood or airway samples, hold promise as non-invasive biomarkers for early diagnosis, risk stratification, and monitoring [48].

3.2. Histone Modifications

Acetylation and methylation of histones change chromatin accessibility and transcription [20]. Histone deacetylase (HDAC) activity is reduced in the lungs of patients with COPD in proportion to the severity of airflow limitation. Given that HDAC down-regulates the production of proinflammatory cytokines, this reduction contributes to the enhanced inflammatory response that characterises COPD [44]. Dysregulation of histone-modifying enzymes (HDACs) and histone acetyltransferases (HATs) contributes to persistent inflammation, immune dysregulation, and apoptosis resistance in COPD, asthma, pulmonary fibrosis, and PAH. Since these changes are reversible, they constitute a promising therapeutic target; several HDAC inhibitors are in preclinical and early clinical development [44].

3.3. Non-Coding RNAs

MicroRNAs (miRNAs) and long non-coding RNAs (lncRNAs) are pivotal post-transcriptional regulators that are themselves epigenetically controlled [44]. Altered miRNA/lncRNA profiles have been associated with immune cell activation, airway remodelling, and chronic inflammation in COPD, asthma, and pulmonary fibrosis [49,50]. Specific miRNA signatures in induced sputum can distinguish smokers with COPD from those without, suggesting their potential as accessible biomarkers [49]. Extracellular vesicles, released upon lung injury, propagate these regulators between cells and thereby amplify disease processes [51].

3.4. Gene–Environment Interactions

The environment moulds the epigenome and determines disease susceptibility [52]. Early-life factors including maternal smoking, pollution, viral infections, and farming environment induce persistent methylation changes that increase later-life risk of asthma, COPD, and PAH [53]. Joubert et al., in a genome-wide consortium meta-analysis, identified numerous CpG sites in newborns associated with maternal smoking during pregnancy, many of which map to genes involved in xenobiotic metabolism and immune regulation [53]. The 17q12–21 locus in asthma represents a strong gene–environment interaction moderated by early-life exposures [20]. Such exposures can reprogram immune responses, epithelial barriers, and tissue repair pathways to shape the origins of these diseases, and these effects may extend beyond classical respiratory genes [54,55].

3.5. Disease-Specific Epigenetic Insights

In COPD, epigenetic regulation of apoptosis, inflammation, and immunity includes methylation of HHIP and FAM13A and dysregulated non-coding RNAs. Genome-wide methylation studies in lung tissue reveal differences related to lung function, nicotine dependence, and T-cell development that may not resolve after smoking cessation [44,56]. In asthma, epigenetic regulation of epithelial and immune genes, notably IL33, TSLP, and the 17q12–21 locus, mediates environmental effects [57]. In ILDs/IPF, telomere length, surfactant biogenesis, and host defence pathways display disease-related epigenetic changes [22]. In PAH, epigenetic modulation of BMP signalling, endothelial function, and vascular remodelling accompanies BMPR2 mutations [58]. In lung cancer, aberrant methylation and chromatin remodelling affect cell cycle, DNA damage, and immune regulation genes. Sandoval et al. identified a prognostic DNA methylation signature for stage I non-small-cell lung cancer (NSCLC) that independently predicted survival [59].

Epigenetic signatures have the potential to provide early biomarkers of disease risk, progression, and therapy response [59]. They are especially useful in young or at-risk populations for detecting preclinical changes [53]. In our view, the pulmonary epigenome represents both a biomarker reservoir and a therapeutic frontier. The key challenge is moving from association to causation: while many epigenetic marks are associated with disease, establishing which are drivers versus passengers requires functional validation through techniques such as epigenome editing and Mendelian randomisation.

4. Molecular Endotyping: From Syndromic to Molecular Classification

4.1. Asthma Endotypes: Type 2-High and Type 2-Low

The molecular endotyping of asthma represents one of the most clinically impactful applications of multi-omic technologies in pulmonary medicine. Asthma is now understood to comprise at least two major molecular endotypes—Type 2 (T2)-high and T2-low—with distinct pathobiological mechanisms, biomarker profiles, and therapeutic implications [16,19].

Type 2-high asthma is characterised by upregulation of Th2 immune pathways, including interleukin (IL)-4, IL-5, and IL-13, along with eosinophilic airway inflammation, often allergic sensitisation, and responsiveness to corticosteroids [16,19]. Epithelial-derived alarmins thymic stromal lymphopoietin (TSLP), IL-33, and IL-25 activate group 2 innate lymphoid cells (ILC2s) and Th2 cells, driving eosinophil recruitment, mucus hypersecretion, and bronchial hyperresponsiveness [21]. Clinical biomarkers of T2 inflammation include blood eosinophils (>300 cells/μL) and fractional exhaled nitric oxide (FeNO > 25 ppb) [16]. Approximately 50% of mild-to-moderate asthma and a larger proportion of severe asthma are T2-high [19].

Type 2-low asthma encompasses neutrophilic asthma (driven by Th1 and Th17 pathways, with IL-17A and TNF-α) and paucigranulocytic asthma (lacking eosinophilic or neutrophilic inflammation) [18,21]. T2-low asthma is poorly responsive to corticosteroids and is not as well characterised, being defined largely by the absence of T2-high markers [18]. It may be associated with obesity, older age of onset, and environmental exposures distinct from those driving T2-high disease.

Peters et al. developed a transcriptomic method using bronchial epithelial brushings to determine airway immune dysfunction in T2-high and T2-low asthma, identifying gene expression signatures that distinguish these endotypes with greater precision than blood biomarkers alone [60]. More recently, Yue et al. demonstrated that nasal epithelial transcriptomic profiles in youth can identify T2-high asthma endotypes non-invasively, with gene expression signatures correlating with blood eosinophils and FeNO [15]. Karp et al. confirmed that nasal gene expression shows a distinct signature in T2-high asthma but not in T2-low disease, suggesting that current transcriptomic tools are more informative for T2-high endotyping [56].

The identification of T2-high endotypes has directly enabled the development and clinical deployment of biologic therapies targeting specific cytokine pathways. Anti-IL-5 agents (mepolizumab, reslizumab, benralizumab) reduce eosinophilic inflammation; dupilumab blocks IL-4 and IL-13 signalling; omalizumab targets IgE; and tezepelumab inhibits TSLP upstream of multiple T2 pathways [19,26]. Itepekimab, a monoclonal antibody targeting IL-33, has shown efficacy in patients with moderate-to-severe asthma with evidence of T2 inflammation, further expanding the therapeutic armamentarium [26]. The 2026 American College of Chest Physicians (CHEST) guideline recommends biologic therapy for adults with severe asthma who remain uncontrolled despite optimised inhaler therapy, with the choice of biologic guided by biomarker profiles including blood eosinophil count, FeNO, and total IgE [27].

4.2. COPD Endotyping Through Multi-Omic Integration

COPD is increasingly being recognised as a syndrome comprising multiple molecular endotypes rather than as a single disease entity [13,14]. Multi-omic approaches are beginning to dissect this heterogeneity with unprecedented resolution. Li et al. integrated transcriptomic, proteomic, and metabolomic data from COPD patients and identified distinct molecular subtypes characterised by the differential activation of inflammatory, metabolic, and tissue-remodelling pathways [52]. Gillenwater et al. developed a multi-omics subtyping pipeline integrating clinical, proteomic, metabolomic, and genomic data, identifying subtypes with distinct inflammatory and metabolic signatures that correlate with clinical outcomes [57]. Zhang et al. used proteomic and metabolomic profiling to reveal panels of circulating diagnostic biomarkers and molecular subtypes in stable COPD, identifying subgroups distinguished by lipid metabolism, oxidative stress, and immune activation patterns [53].

Olvera et al. performed lung tissue multilayer network analysis integrating transcriptomic, epigenomic, and genomic data, uncovering the molecular heterogeneity of COPD at the tissue level [51]. This approach identified disease-relevant modules enriched in immune signalling, extracellular matrix remodelling, and mitochondrial dysfunction pathways, providing a systems-level view of COPD pathobiology that cannot be captured by any single omic layer alone. Moll et al. developed a blood-based transcriptional risk score (TRS) composed of more than 100 transcripts that predicts COPD susceptibility and FEV1 decline [19]. Incorporation of the TRS with the PRS further improves prognostic precision, suggesting that combining genetic and transcriptomic information captures complementary dimensions of disease risk [19]. Blood eosinophil count has emerged as one validated biomarker predicting response to inhaled corticosteroids and, more recently, to anti-IL-5 biologics in COPD, but additional biomarkers are needed to capture the full spectrum of COPD heterogeneity [54,61]. Figure 4 illustrates the multi-omic endotyping frameworks across respiratory diseases.

4.3. Multi-Omic Endotyping in Interstitial Lung Diseases

In ILDs, multi-omic approaches are advancing both diagnostic classification and prognostic stratification. The Envisia Genomic Classifier (Veracyte) uses a 190-gene machine learning classifier applied to transbronchial biopsy samples to detect a molecular usual interstitial pneumonia (UIP) signature [19,62]. A systematic review and meta-analysis demonstrated that the genomic classifier achieves a specificity of 92% (95% CI: 81–95%) and a sensitivity of 68% (95% CI: 55–73%) for predicting histopathological UIP [63]. When combined with high-resolution computed tomography (HRCT) patterns, the classifier increased sensitivity to 79.2% while maintaining specificity at 90.6% [19]. The 2025 ERS/ATS update on the classification of interstitial pneumonias acknowledges the Envisia classifier as accessible in North America and Europe, noting that it strengthens UIP diagnosis but that its clinical utility remains unproven and it cannot differentiate among different causes of UIP or among various non-UIP patterns [62].

Kheir et al. demonstrated that the genomic classifier increased diagnostic confidence in multidisciplinary discussions (MDDs) from 43% to 93% in patients with probable UIP when added to bronchoscopic lung cryobiopsy, with concordance coefficients of 0.92 between the classifier and MDD diagnosis [64]. Beyond the Envisia classifier, peripheral blood transcriptomic signatures are emerging as prognostic tools, with serum protein biomarkers including KL-6, MMP-7, CCL18, and SP-D being evaluated for prognostic stratification in fibrosing ILDs [54,62].

5. Molecular Diagnostics: From Syndromic to Genomic Classification

5.1. Nucleic Acid Amplification Tests and Multiplex Panels

Molecular diagnostics have transformed respiratory disease detection by providing the rapid, sensitive identification of pathogens and molecular signatures. Nucleic acid amplification tests (NAATs), including real-time reverse transcription polymerase chain reaction (RT-PCR) and isothermal amplification, have become predominant over culture-based and antigen assays for community- and hospital-acquired infections [65]. These technologies detect a wide range of viral and bacterial pathogens, including Influenza A/B, respiratory syncytial virus (RSV), SARS-CoV-2, parainfluenza, adenovirus, human metapneumovirus, and atypical bacteria like Mycoplasma pneumoniae, from different respiratory specimens within hours [66]. Multiplex molecular panels allow for the simultaneous detection of multiple organisms in a single run, improving yield and facilitating appropriate antimicrobial stewardship [67]. Point-of-care molecular tests, many of which are CLIA-waived, hold particular value for early antiviral initiation, infection control, and the management of outbreaks [68].

5.2. Next-Generation Sequencing and Metagenomics

Next-generation sequencing (NGS) technologies, especially metagenomic NGS (mNGS), allow for the unbiased detection of all possible pathogens in a sample [69]. In contrast to targeted PCR, mNGS detects bacteria, viruses, fungi, and parasites without prior hypothesis, which is extremely useful in difficult-to-diagnose cases [36].

Multiple studies have now established the clinical utility of mNGS in pulmonary infections. A multicentre retrospective study of 246 patients with suspected pulmonary infection demonstrated that mNGS had significantly higher sensitivity than conventional testing (53.49% vs. 23.26%), with particular advantages for Mycobacterium tuberculosis (67.86% vs. 17.86%), atypical pathogens (100% vs. 7.14%), viruses (92.31% vs. 7.69%), and fungi (78.57% vs. 39.29%) [36]. A larger study of 400 patients demonstrated an mNGS sensitivity of 93.3% versus 55.6% for culture, with an AUC of 0.744 versus 0.636 [67]. Importantly, mNGS detected organisms frequently missed by culture, including Streptococcus pneumoniae, Haemophilus influenzae, Aspergillus species, and Pneumocystis jirovecii [67]. mNGS results led to treatment modifications in 50–72% of patients across studies, including both escalation and de-escalation of antimicrobial therapy [67,70]. In our view, mNGS is most valuable in immunocompromised patients, critically ill patients with culture-negative infections, and cases where mixed or rare pathogen infections are suspected [33,71,72].

5.3. Liquid Biopsy and Tumour Genomics in Lung Cancer

Liquid biopsy represents the analysis of tumour-derived material, including circulating tumour DNA (ctDNA), cell-free DNA (cfDNA), circulating tumour cells (CTCs), and exosomes, from blood or other body fluids [38,73]. This minimally invasive method enables the real-time genomic profiling of lung cancers, identification of actionable driver mutations, and monitoring for minimal residual disease (MRD) or emerging resistance [35,74].

In advanced NSCLC, ctDNA testing is used to identify actionable genomic alterations when tissue is unavailable or insufficient. The sensitivity of detecting target mutations with ctDNA is 60–80%, depending on the tumour location, size, and vascularity, as well as the detection method [75]. The American Society of Clinical Oncology (ASCO) provisional clinical opinion states that in patients without tissue-based genomic test results, treatment may be based on actionable alterations identified in cfDNA, and that cfDNA testing is most helpful when archival tissue is unavailable and new tumour biopsies are not feasible [76]. FDA-approved plasma tests currently include detection of EGFR mutations, with osimertinib approved for EGFR T790M mutation-positive NSCLC based on cfDNA testing [66,75].

A systematic review and meta-analysis of 27 studies (2424 patients) demonstrated that baseline ctDNA-negative patients had significantly higher progression-free survival (pooled HR: 2.97; 95% CI: 1.92–4.85) and overall survival (pooled HR: 3.49; 95% CI: 1.73–6.95) compared with ctDNA-positive patients [28]. Early reduction in or clearance of ctDNA levels after treatment further improved PFSs (pooled HR: 3.78; 95% CI: 1.91–7.38) and OS (pooled HR: 2.20; 95% CI: 1.49–3.28) [28]. In early-stage NSCLC, ctDNA-based MRD detection is emerging as a tool for guiding adjuvant therapy decisions. Patients with preoperative ctDNA-negative status have demonstrated improved 5-year overall survival compared with ctDNA-high patients (100% vs. 48.8%) [77]. Beyond ctDNA, the scope of liquid biopsy is broadening to include circulating RNAs, exosomes, DNA methylation signatures, and tumour-educated platelets, each providing complementary insights into tumour biology [35,48,78].

5.4. Genomic Classifiers in Interstitial Lung Diseases

As discussed in Section 4.3, the Envisia Genomic Classifier applied to transbronchial biopsy samples can distinguish UIP from non-UIP histopathological patterns with high specificity (92%) and moderate sensitivity (68%) [19,63]. The 2025 ERS/ATS classification update recognises this technology as a diagnostic adjunct while noting that it cannot differentiate causes of UIP or classify non-UIP patterns [62].

5.5. Monogenic Respiratory Disorders

In cystic fibrosis, modern newborn-screening protocols utilise immunoreactive trypsinogen measurement followed by DNA analysis for common and ethnic-specific mutations. High-throughput sequencing achieves >99% sensitivity and directs genotype-specific therapy with CFTR modulators [29,32]. Primary ciliary dyskinesia (PCD) is caused by biallelic pathogenic variants in more than 55 genes that impact the structure or function of motile cilia [79]. Sequencing panels now identify up to 70% of cases [42]. In AATD, serum level measurement, targeted genotyping, or sequencing enables definitive diagnosis and eligibility for augmentation therapy [80]. For PAH, targeted sequencing panels that include BMPR2, SMAD9, CAV1, KCNK3, and related genes confirm heritable disease and inform prognosis and family screening [25,26].

6. Prognostic Biomarkers in Pulmonary Disease

6.1. Polygenic Risk Scores

As detailed in Section 2.2 and Section 2.3, the PRS aggregates millions of variants to quantify inherited disease risk [41]. Large meta-analyses demonstrate that PRSs derived for lung function and asthma are strongly associated with COPD, asthma, and asthma–COPD overlap, with an odds ratio up to approximately 8 in the highest decile [7]. While odds ratios provide a useful measure of relative genetic risk, they should be interpreted cautiously, as they may overestimate absolute risk differences, particularly for common outcomes. Risk ratios or absolute risk differences, where available, may provide more clinically intuitive measures of effect size. The PRS also predicts earlier onset, emphysema subtypes, and impaired lung growth. Combining the PRS with clinical factors such as age, sex, and smoking increases predictive accuracy, with an AUC of approximately 0.80 [28,48].

6.2. Transcriptomic and Proteomic Signatures

Blood-based transcriptional risk scores (TRSs) composed of more than 100 transcripts predict COPD susceptibility and FEV1 decline [19]. Incorporation of the TRS with the PRS further improves prognostic precision [19]. Proteomic and metabolomic biomarkers reflecting inflammation (CRP, IL-6), oxidative stress (8-isoprostane), and immune activation provide accessible measures for disease monitoring and mortality prediction [61,64]. Reproducible circulating protein markers identified in proteomic studies include surfactant protein D (SP-D), club cell secretory protein (CC16), and fibrinogen, which predict disease severity, exacerbation risk, and mortality in COPD [61]. Beyond individual omic layers, integration of genomics, transcriptomics, proteomics, metabolomics, and microbiomics defines molecular endotypes with distinct inflammatory, metabolic, and microbial characteristics. Multi-omic clustering in COPD identifies subgroups with distinct immune response and metabolic pathways, and pathway enrichment consistently identifies immune-related and nicotinic acetylcholine receptor signalling pathways as central in disease processes.

6.3. Epigenetic and MicroRNA Biomarkers

Genetic and environmental influences on respiratory diseases are integrated by epigenetic modifications and non-coding RNAs [62]. MicroRNAs can distinguish between asthma and COPD endotypes, predict steroid responsiveness, and act as stable non-invasive biomarkers [49]. Circulating miR-15b-5p has been identified as a biomarker for asthma–COPD overlap, demonstrating the potential for miRNAs to aid in the classification of overlapping airway diseases [49]. DNA methylation profiles offer potential for early diagnosis and reflect disease activity [62].

6.4. Prognostic Biomarkers in Interstitial Lung Disease

Biomarkers including KL-6, MMP-7, and CCL18 and genetic markers of epithelial dysfunction are promising in predicting disease progression and response to therapy in fibrosing ILDs [54,62]. Telomere length measurement is emerging as a prognostic and potentially therapeutic biomarker, with short telomeres associated with worse outcomes and potentially guiding post-transplant immunosuppression strategies [62].

6.5. Combined Clinical–Molecular Risk Models

Combining biological-ageing indices (e.g., PhenoAgeAccel) with genetic risk provides improved prediction of the onset of COPD, IPF, and asthma [28,81]. This captures environmental mediation through smoking and air pollution, linking molecular ageing to clinical outcomes.

Table 2. Summary of diagnostic and prognostic biomarkers across respiratory diseases.

Biomarker Type	Specific Biomarkers	Disease Application	Clinical Utility (Diagnostic/Prognostic)	Evidence Level
Polygenic Risk Score (PRS)	COPD PRS (279 variants); asthma PRS (lasso/Bayesian)	COPD; asthma; ACO	Diagnostic (risk stratification); prognostic (disease severity)	Level IIa–IIb; meta-analyses
Transcriptomic Signatures	Blood TRS (>100 transcripts); nasal T2-high signature; bronchial epithelial score	COPD; asthma; ILD	Prognostic (FEV1 decline); diagnostic (endotyping)	Level IIb; prospective cohorts
Proteomic Markers	SP-D, CC16, fibrinogen, KL-6, MMP-7, CCL18, IL-6, CRP	COPD; IPF; PAH; lung cancer	Prognostic (severity, exacerbations, mortality)	Level IIa; multi-cohort studies
Epigenetic Markers	HHIP/FAM13A methylation; CYP1B1 methylation; stage I NSCLC methylation signature	COPD; asthma; lung cancer	Diagnostic; prognostic; therapeutic monitoring	Level IIb–III; replication needed
MicroRNA Biomarkers	miR-15b-5p (ACO); sputum miRNA panels (COPD vs. non-COPD)	COPD; asthma; ACO	Diagnostic (endotyping); predictive (steroid response)	Level IIb–III; validation ongoing
Composite Clinical–Molecular	PRS + age/sex/smoking (COPD AUC ≈0.80); PhenoAgeAccel + PRS	COPD; asthma; IPF	Diagnostic (superior to single markers)	Level IIa; prospective validation needed
ctDNA/Liquid Biopsy	Baseline ctDNA positivity; ctDNA clearance; EGFR T790M (plasma)	Lung cancer (NSCLC)	Prognostic; predictive (therapy response); MRD monitoring	Level IIa; RCT data emerging

ACO—asthma–COPD overlap; AUC—area under the curve; CC16—club cell secretory protein; COPD—chronic obstructive pulmonary disease; CRP—C-reactive protein; ctDNA—circulating tumour DNA; FEV1—forced expiratory volume in 1 s; IPF—idiopathic pulmonary fibrosis; MRD—minimal residual disease; NSCLC—non-small-cell lung cancer; PAH—pulmonary arterial hypertension; PRS—polygenic risk score; RCT—randomised controlled trial; SP-D—surfactant protein D; TRS—transcriptional risk score.

summarizes diagnostic and prognostic biomarkers across respiratory diseases.

7. Precision Medicine in Respiratory Diseases

7.1. Genomic Foundations of Precision Respiratory Medicine

Extensive GWASs have identified thousands of variants that influence asthma, COPD, and ILD susceptibility and phenotypes. Loci near IL33, TSLP, and other immune genes drive asthma endotypes, while HHIP, FAM13A, and SERPINA1 variants shape COPD subtypes. These findings inform risk prediction and have already enabled genotype-directed therapies: CFTR modulators in CF and AAT augmentation in AATD represent the most mature examples [29,32,82].

7.2. The Treatable-Traits Framework

The concept of treatable traits represents a paradigm shift in the management of chronic airway diseases, moving beyond traditional diagnostic labels (asthma, COPD) to a precision medicine framework that identifies and targets modifiable characteristics in each individual patient [58,83]. A treatable trait is defined as a recognisable phenotypic or endotypic characteristic that can be assessed and successfully targeted by therapy to improve a clinical outcome [83]. Traits are identified across three domains: pulmonary (e.g., eosinophilic airway inflammation, airflow limitation, mucus hypersecretion), extrapulmonary (e.g., obesity, gastroesophageal reflux, cardiovascular comorbidity, anxiety/depression), and behavioural/risk factor (e.g., smoking, poor inhaler technique, medication non-adherence) [58,83].

Biomarkers are central to the treatable-traits paradigm, translating latent endotypes into measurable traits that inform treatment selection [54]. Blood eosinophil count (BEC) predicts the benefit of inhaled corticosteroids in both asthma and COPD; FeNO indicates steroid responsiveness in asthma; and sputum cell profiles enable identification of neutrophilic or eosinophilic airway inflammation [54,84]. Quantitative CT metrics extend trait definition to structural and functional domains, while multi-omic and microbiome signatures reveal the molecular endotypes that underpin disease heterogeneity [54]. Trials have demonstrated the superiority of the treatable-traits approach, with significant improvements in asthma control and quality of life compared with usual care [7]. We view the treatable-traits framework as the most practical bridge between multi-omic discovery and clinical implementation in respiratory medicine.

7.3. Biologic Therapies Guided by Molecular Endotyping

The development of biologic therapies for severe asthma exemplifies the successful translation of molecular endotyping into precision therapeutics. Currently approved biologics target specific nodes in the T2 inflammatory cascade: anti-IgE (omalizumab) for allergic asthma; anti-IL-5/IL-5R agents (mepolizumab, reslizumab, benralizumab) for severe eosinophilic asthma; anti-IL-4Rα (dupilumab) blocking both IL-4 and IL-13 signalling; and anti-TSLP (tezepelumab) targeting the upstream alarmin TSLP [19,26]. The 2026 CHEST guideline recommends biologic therapy for adults with severe asthma who remain uncontrolled despite optimised inhaler therapy, with the choice of biologic guided by biomarker profiles [27]. Couillard et al. have proposed a biomarker-guided algorithm for choosing the right biologic for the right patient, incorporating BEC, FeNO, total IgE, and clinical phenotype into a decision framework [28].

In COPD, dupilumab has recently demonstrated efficacy in reducing exacerbations in patients with evidence of T2 inflammation (blood eosinophils ≥ 300 cells/μL), representing the first biologic therapy approved for COPD and extending the precision medicine paradigm from asthma into obstructive lung disease more broadly [14,85,86].

8. Challenges and Future Directions

8.1. Functional Interpretation of Genetic Variants

The interpretation of the vast variant landscape from GWASs and sequencing remains complex: many variants reside in non-coding regions with unclear function. Moving from statistical association to biological mechanism requires functional genomic approaches, including CRISPR-based perturbation screens, massively parallel reporter assays, and integration with chromatin accessibility and gene expression data. The post-GWAS era demands systematic efforts to connect genetic signals to causal genes, cell types, and pathways.

8.2. Data Integration and Bioinformatics

The integration of large multi-omic datasets requires robust bioinformatics infrastructure and standardisation [87]. Current challenges include heterogeneity in sample processing, platform variability, batch effects, and the lack of standardised analytical pipelines [88]. Dimensionality reduction, multi-omic association methods, and machine learning approaches are being developed to address these challenges, but validation and reproducibility remain concerns [89].

8.3. Diversity and Equity

A major gap in diversity exists. Most genomic data are derived from European-ancestry cohorts, reducing global applicability and potentially exacerbating health disparities [82]. Martin et al. have demonstrated that the clinical use of the current PRS may exacerbate health disparities because of differential predictive performances across ancestries [82]. Expanding multi-ancestry participation in GWASs, biobanks, and clinical trials is essential to ensure equitable precision medicine. Approaches such as polygenic transcriptome risk scores that improve cross-ethnic portability represent promising solutions [90]. Beyond genomic equity, disparities in access to care compound the challenges of translating precision medicine into equitable outcomes [91,92].

8.4. Clinical Translation

Clinical translation lags behind discovery. Whereas genotyping for monogenic diseases such as AATD and CF has become routine, application to polygenic disorders like asthma and COPD is limited [66,93]. Cost, lengthy turnaround times, and interpretive complexity are formidable barriers to adoption, especially in resource-constrained settings [94]. The development of point-of-care biomarker assays, simplified risk scores, and clinical decision support tools will be essential for bridging the gap between discovery and practice.

8.5. Emerging Frontiers

8.5.1. Single-Cell and Spatial Omics

Single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics are revealing unprecedented cellular heterogeneity in the lung, identifying novel cell states in health and disease [95,96]. The Human Lung Cell Atlas, generated from approximately 75,000 human cells, has defined the gene expression profiles and anatomical locations of 58 cell populations, including 14 previously unknown cell types [96]. Firsova et al. created a spatially resolved tissue reference atlas of the healthy human lung and demonstrated previously unknown imbalances of epithelial cell type compositions in COPD lungs [97]. Spatial transcriptomics has delineated distinct tumour microenvironments in lung cancer, revealing prognostically significant gene signatures [98]. In IPF, spatial approaches have mapped the ecosystem of immune cells, inflammatory mediators, and fibroblast subtypes within discrete regions of diseased tissue, offering mechanistic insights into disease progression and tissue remodelling [98,99].

8.5.2. The Respiratory Microbiome

Metagenomic approaches have substantially advanced understanding of host–microbe interactions in respiratory disease [100,101]. In asthma, early-life microbial perturbations, reduced community diversity, and enrichment of Streptococcus and Moraxella are associated with disease development, while gut dysbiosis influences immune maturation and tolerance through the gut–lung axis [100,102]. In COPD, adult-onset dysbiosis with Proteobacteria dominance and depletion of commensal anaerobes such as Prevotella and Veillonella are characteristic features [100,103]. Xue et al. demonstrated that the sputum microbiome is associated with disease status in COPD and can distinguish different inflammatory endotypes: Gemella was associated with Th2 inflammatory endotypes, whereas Prevotella was associated with Th17 inflammatory endotypes [104]. Different airway inflammatory phenotypes (eosinophilic vs. non-eosinophilic) correlate with specific fungal and bacterial microbiota in both asthma and COPD [105].

8.5.3. Artificial Intelligence and Machine Learning

AI and machine learning are increasingly being integrated across the spectrum of respiratory genomics and clinical care [106,107]. In diagnostics, deep learning systems streamline chest-imaging workflows, classify lung nodules on CT, and detect technical errors in pulmonary function testing [108]. In genomics, machine learning algorithms are being applied to multi-omic data integration, enabling dynamic phenotyping, prediction of treatment response, and identification of novel disease subtypes [106,109]. AI-driven drug discovery has already identified novel therapeutic targets for IPF, including an NCK-interacting kinase inhibitor with strong antifibrotic properties [109]. In chronic care, connected devices integrated with environmental data help forecast asthma and COPD exacerbations, while telehealth and predictive models enable earlier, more personalised interventions [108].

8.5.4. Genomics-Guided Drug Discovery and Repurposing

The functional interpretation of GWAS signals is opening new avenues for drug target identification and repurposing [98,110]. El-Husseini et al. identified 128 independent SNPs associated with asthma and proposed 161 possible drug targets through functional interpretation, of which 16 are already targeted by existing asthma drugs [110]. Novel cell surface receptor targets include IL7R, CCR7, IL2RB, GPR183, and ITGB8 [110]. Pathway enrichment analyses across chronic respiratory diseases consistently identify immune-related processes and nicotinic acetylcholine receptor pathways as enriched, providing a biological rationale for targeted drug development [99]. Mendelian randomisation, transcriptome-wide association studies, and polygenic scoring are being used to prioritise drug repurposing candidates, with causal evidence supporting their therapeutic potential [98].

8.5.5. Pharmacogenomics

Beyond drug target discovery, pharmacogenomics aims to predict individual treatment response based on genetic variation [100,101]. In asthma, polymorphisms in ADRB2 (encoding the β2-adrenergic receptor), ALOX5 (encoding 5-lipoxygenase), and glucocorticoid receptor pathway genes have been associated with variable responses to bronchodilators, leukotriene modifiers, and corticosteroids, respectively [100,101]. In CF, the paradigm of genotype-directed CFTR modulator therapy demonstrates the full potential of pharmacogenomics when a clear genotype–drug relationship exists [29,32]. For polygenic diseases, pharmacogenomic prediction is more complex, but the integration of the PRS with pharmacogenomic markers may eventually enable more precise treatment selection.

9. Conclusions

The genomic revolution has profoundly advanced the understanding of chronic respiratory diseases. High-throughput genomics, sequencing, and multi-omic integration have identified numerous loci that confer susceptibility, key regulatory pathways, and therapeutic targets across asthma, COPD, ILDs, PAH, CF, and lung cancer. The development of polygenic risk scores now enables quantitative risk stratification, with the COPD PRS achieving up to 8-fold increased odds in the highest-risk decile and combined clinical–genetic models approaching AUCs of 0.80. Molecular endotyping, most maturely exemplified by T2-high/T2-low classification in asthma, has directly enabled the deployment of targeted biologic therapies, transforming outcomes for patients with severe disease. The success of genotype-directed CFTR modulator therapy in CF provides a compelling proof of concept for precision medicine in pulmonary disease.

However, significant challenges remain. Functional interpretation of the vast majority of GWAS-identified variants is incomplete. Most genomic data derive from European-ancestry populations, limiting the global applicability of current risk prediction tools and potentially exacerbating health disparities. The integration of massive multi-omic datasets requires robust bioinformatics infrastructure and standardised analytical frameworks that are not yet widely available. Clinical translation of molecular discoveries into routine practice is hampered by cost, complexity, and the need for prospective validation.

Emerging frontiers including single-cell and spatial transcriptomics, respiratory microbiome profiling, artificial intelligence, and genomics-guided drug discovery promise to accelerate the pace of discovery and translation. The treatable-traits framework offers a practical bridge between molecular complexity and clinical decision making, decomposing heterogeneous diseases into modifiable components amenable to targeted intervention. Continued investment in functional genomics, integrative analytics, diverse population representation, and equitable access to genomic technologies is essential to maximise the potential of genomics in pulmonary medicine. Ultimately, these advances will transform prevention, diagnosis, and treatment, ushering in a new era of personalised respiratory health and precision medicine.