DNA was extracted as previously described (10 (link), 11 (link)). DNA was sequenced using an AmpliSeq Custom Panel designed to amplify the coding regions of CPA1 and CPB1. NGS was performed with 540 chips (Ion S5 system, single-read, 200 bp read-length) and reads from the Ion Torrent Server were mapped (HG19) and variants called using NextGENe Software (Softgenetics, LLC, State College, PA) according to manufacturer’s protocols as previously described (38 (link)). Median depth of coverage for CPA1 and CPB1 reads was 262 reads per amplicon. A minimum percentage coverage of reads of ≥10 was required for at least 85% of amplicons. Samples below this coverage were re-sequenced to obtain sufficient coverage. An amplicon depth of coverage of 15 reads was required to call a variant, and candidate variant reads of unknown significance were visualized with Integrated Genomics Viewer (IGV) software. Candidate variants of unknown significance were confirmed by Sanger sequencing, performed at the Johns Hopkins DNA sequencing core.