Results

SARP participants with asthma (n= 1401, SARP1-3) had whole genome sequencing(after consent was attained) through the NHLBI-sponsored TOPMed program. We extracted sequencing data on functional variation spanning a 190.5Kb region of CFTR (hg19 position chromosome 7:117119040-117309560). Potentially pathogenic, low frequency-to-rareCFTR variants (allele frequency<0.05) were classified as (1) CF-causing based on the CFTR2 database (cftr2.org), (2) having varying clinical consequences on CF risk (VVCC) based on CFTR2, and (3) likely pathogenic based on the American College of Medical Genetics (ACMG)(acmg.net) assuming a recessive model (Figure 1 ). The minor allele frequency of these potentially pathogenic CFTR variants and regression-based models were stratified by self-reported racial/ethnic group (African American and non-Hispanic whites). Regression-based models tested for associations of potentially pathogenic genotypes with clinical outcomes of interest including the comparison between (1) individuals with no identifiable potentially pathogenic variation versus 1 potentially pathogenic variant, (2) one or more variants, or (3) 2 potentially pathogenic variants using collapsing-based burden tests. Variant-specific models also compared individuals with and without F508del and no other potentially pathogenic variants (i.e. F508del heterozygotes).
Of 1401 total participants, 9.5% (134) were carriers of one potentially pathogenic CFTR variant and these carriers were more likely to be non-Hispanic white (NHW, 10.1% [84 of 831]) when compared to African American individuals (AA, 5.2% [22 of 426]). The most frequent potentially pathogenic CFTR variant identified was F508del, found in 30 individuals, the majority of whom were NHW (3.5% [29 of 831]). We found ≥2 potentially pathogenic CFTR  variants in 1.4% (19) of total participants which occurred more frequently in African American (2.8%, n=12) compared to NHW individuals (0.5%, n= 4 ). Potentially pathogenicCFTR variant genotypes (none versus ≥1 or ≥2 potentially pathogenic variants), including F508del, were not cumulatively associated with lung function measures or exacerbations requiring corticosteroid bursts, ED visits, or hospitalizations. In four NHW individuals, we found three with F508del compound heterozygosity with a VVCC: two c.1210-11T>G(legacy: 5T;TG12), and one with c.3208C>T(p.Arg1070Trp; legacy: R1070W; a CFTR modulator therapy-eligible variant) and an individual homozygous for the VVCC 5T;TG12. All four individuals were females with a reduced pre-bronchodilator forced expiratory volume in 1 second (FEV1) percentage predicted (Table 1 ). All 12 AA with two potentially pathogenic variants were carriers of the cis-variants c.220C>T (p.Arg74Trp; legacy: R74W) and c.3808G>A (p.Asp1270Asn; legacy: D1270N), both FDA-approved targets for CFTR modulator therapy and highly likely to be occurring in cis and in full linkage disequilibrium. The CF-causing variants we identified in NHW and AA, especially F508del, had a markedly higher minor allele frequencies (MAF) in SARP participants compared to reference general populations (Table 2).