To the Editor: The comparison of CFTR mutation screening panels with population frequencies using data from the Exome Aggregation Consortium carried out by Lim and colleagues1 highlights the continued gap in annotation of variation in CFTR. The Exome Aggregation Consortium cohort presents a much more ethnically and geographically comprehensive sampling of individuals than the collections of cystic fibrosis (CF) patients from which mutation screening panels are derived. The authors therefore conclude that the sensitivity of CF screening could be improved by replacing panels with exome sequencing.

Sequencing has an important role in CF patients with defined phenotypes and will enhance understanding of the role of CFTR variants in non-European individuals who present with phenotypic features not readily recognized as CF.

However, we feel that sequencing is not a better way to achieve the goals of carrier screening. Carrier screening has traditionally focused on variants that have known, life-threatening consequences in order to enable unaffected heterozygous patients to make informed reproductive decisions.2 Sequencing and the potential identification in screened individuals of uncharacterized variants that may have reduced penetrance promotes selection against relatively benign phenotypes along with true disease-causing conditions. Furthermore, the use of computational mutation prediction scores from PolyPhen-2 and PROVEAN are too imprecise and undervalidated for clinical decision making because they lack specificity.3 For example, a variant with a slight effect on reproductive fitness (such as those associated with only obstructive azoospermia) is categorized as pathogenic in the same fashion as a variant that is fully penetrant for life-shortening CF. Therefore, the authors may be overestimating the number of CF-causing mutations that go undetected on traditional carrier screens. Indeed, the lack of information to adequately counsel an asymptomatic carrier with an uncharacterized variant is an important unmet challenge in implementing personalized medicine.4

Disclosure

The authors declare no conflict of interest.