Whole-Genome Sequencing in Common Respiratory Diseases. Ready, Set, Go!

High-throughput (next-generation) sequencing technologies are becoming increasingly used in studies of common, complex diseases. As with many genomic technologies, studies in cancer have led the way, through projects such as The Cancer Genome Atlas (1). However, these sequencing techniques are becoming more widespread in benign diseases, including respiratory disorders such as asthma, chronic obstructive pulmonary disease (COPD), pulmonary fibrosis, and pulmonary hypertension (2–5). RNA sequencing has several advantages over microarrays, such as the ability to assay greater number of RNA species with increased sensitivity. Costs are equivalent, if not favorable, for RNA sequencing compared with microarrays.

In the very near future, DNA sequencing is poised to replace genotyping arrays, which have been the standard in genome-wide association studies (GWASs) for more than a decade. Sequencing of the exome, defined as the protein-coding portion of the genome, has served as an intermediate step, as costs were initially too high for whole-genome sequencing (WGS) in large sample sizes. Exome sequencing has been quite fruitful in finding causal mutations in rare Mendelian diseases (6), which are frequently due to nonsynonymous variants that lead to amino acid substitutions. Exome sequencing has not been quite as successful in complex diseases, which are more frequently influenced by variants outside of the coding genome, which likely affect gene regulation (7). WGS is able to overcome this limitation, and costs have decreased to the point where it is now feasible to perform WGS in the sample sizes necessary to find rare variants associated with common diseases.

In this issue of the Journal, Radder and colleagues (pp. 159–171) present the first reported WGS study in COPD (8). The investigators used the strategy of studying individuals with extreme phenotypes, a strategy that has been used in many prior genetic studies, including studies of COPD (3). From a Pittsburgh cohort of current or former smokers, they selected cases with both emphysema on the basis of quantitative chest computed tomography scans and severe airflow obstruction (FEV₁ < 40% predicted), who were compared with control subjects without emphysema or airflow obstruction, matched by age, sex, and smoking history. This yielded a modest sample size of 65 susceptible and 64 resistant smokers with high-quality WGS data. WGS resulted in more than 13 million single-nucleotide polymorphisms across all samples. The subsequent analyses focused on the 5.5 million rare variants in the cohort (allele frequency < 1%), an average of approximately 80,000 rare variants per subject. In the analyses of individual variants and groups of sequential variants across the genome (within 30-kb windows), none of the findings reached traditional levels of genome-wide significance. Nor were statistically significant results found in an analysis restricted to rare variants located within introns, exons, and 1-kb flanking regions of genes. However, suggestive associations with several genes were identified through this series of tests, including ZNF816 (zinc finger protein 816) and CCDC38 (coiled-coil domain–containing 38), the latter of which was previously found in an exome sequencing study of resistant smokers with normal lung function (9).

The most interesting result came from an analysis of nonsynonymous coding variants. In particular, a rare nonsynonymous substitution (rs61754411) in the gene PTPRO (protein tyrosine phosphatase, receptor type O) occurred in eight susceptible individuals but was not found in the resistant population. Though the top result, this association did not reach genome-wide significance. The investigators were able to validate this finding by genotyping this variant in the overall Pittsburgh cohort of 686 individuals, demonstrating associations with both FEV₁ and percent emphysema on chest computed tomography scans. Primary human bronchial epithelial cells harboring the mutant allele in PTPRO demonstrated reduced epidermal growth factor signaling in response to ligand or exposure to cigarette smoke extract. Although these are preliminary findings, alterations in epidermal growth factor signaling have been previously implicated in the pathogenesis of COPD (10), supporting the possibility that PTPRO may represent a novel COPD susceptibility gene. The results are promising, yet additional work is needed to confirm the association of this variant with COPD and emphysema and to further elucidate the role of PTPRO in disease pathogenesis.

Although the study by Radder and colleagues (8) is one of the first WGS studies in any common respiratory disease, it follows a considerable history of successful GWAS studies. Some of the lessons learned in the GWAS era of human genetics are immediately applicable to WGS studies, especially the need for large sample sizes to find significant genetic associations in heterogeneous conditions such as COPD. To achieve these sample sizes often requires national or international consortia. Studies of type 2 diabetes (11), serum lipid levels (12, 13), and bone mineral density (14) have incorporated WGS in thousands of subjects. Although sequence alignment and variant detection methods are becoming more standardized (15), there remain many unanswered questions related to WGS study design and interpretation, such as how to choose the optimal data analysis method(s) and even how to define a threshold for statistical significance (16). Replication in additional cohorts and functional validation of identified variants, both of which were undertaken by Radder and colleagues (8), are important steps in any genetic study and are clearly relevant to WGS studies going forward.

The NHLBI-sponsored Trans-Omics for Precision Medicine (TOPMed) project is currently performing WGS in tens of thousands of subjects across multiple heart, lung, blood, and sleep disorders, including asthma, COPD, sleep apnea, sarcoidosis, and pulmonary hypertension, along with quantitative measures of lung function (https://www.nhlbiwgs.org). These studies will surely identify novel genetic risk factors for these common diseases. Moreover, the data will be made publically available through the NCBI database of Genotypes and Phenotypes (dbGaP, https://www.ncbi.nlm.nih.gov/gap/), allowing investigators to perform additional studies as well as to develop and refine WGS data analysis methods. The next generation of respiratory genetics researchers needs to be prepared to handle this deluge of data to identify novel disease genes and targets for intervention.