CYP2D6 genotype analysis

CYP2D6 haplotypes (commonly referred to as ‘star’ alleles) are designated by an asterisk and a combination of roman letters and Arabic numerals, as defined by the Human Cytochrome P450 (CYP) Allele Nomenclature Database at www.cypalleles.ki.se/.³ Of note, the CYP2D6 haplotype in the reference genome (GRCh37) corresponds to CYP2D6*2 and not CYP2D6*1, which is regarded as the wild-type or reference sequence in the pharmacogenomic community.

CYP2D6 genotyping was performed as described.^37,39–44 Briefly, long-range PCR was used to amplify a 6.6-kb fragment encompassing the entire CYP2D6 gene (fragment A), a 3.5-kb fragment from the intergenic region of CYP2D6 duplication structures (fragment B) and a 5-kb fragment from CYP2D7/2D6 hybrid structures (fragment H).⁴² Presence of fragments was determined by band visualisation following agarose gel electrophoresis. The gene regions amplified are shown in Supplementary Figure 1.

To test for single-nucleotide variations, XL-PCR amplicons were diluted 2,000-fold and used in TaqMan genotyping assays (Thermo Fisher Scientific, Waltham, MA, USA) to detect a panel of CYP2D6 ({"type":"entrez-nucleotide","attrs":{"text":"NM_000106.5","term_id":"392513720","term_text":"NM_000106.5"}}NM_000106.5) sequence variations allowing us to assign haplotypes defined as CYP2D6*2, *3, *4, *6, *7, *9, *10, *11, *17, *29, *31, *35, *41, *42 and *45. In the absence of these variants, the haplotype assigned was CYP2D6*1. CYP2D6 duplications/multiplications, the CYP2D6*5 gene deletion, CYP2D7/2D6 hybrid arrangements (collated under the CYP2D6*13 designation⁴⁵), and other CYP2D6/2D7 hybrids (such as CYP2D6*68), were identified by a quantitative CNV assay and confirmed by long-range PCR.^37,44 Furthermore, duplicated gene copies were genotyped by performing TaqMan genotyping assays on an XL-PCR product (fragment D) that encompasses the entire duplicated gene copy. An overview of the genotyping strategy and additional details are provided in Supplementary Table 1.

An Activity Score (AS) was assigned to each allele as described previously⁴⁴ with the traditional phenotype classifications poor, intermediate, extensive and ultrarapid metabolisers in accordance with guidelines from the Clinical Pharmacogenetics Implementation Consortium.^9–11

The CYP2D6 locus, including at least 600 and 150 nucleotides upstream and downstream of the translation start and stop codons, respectively, was sequenced in both directions. The 6.6-kb CYP2D6 fragment A (Supplementary Figure 1) was purified with a GenElute PCR Clean-up kit (Sigma, St Louis, MO, USA). Sequencing was performed with BigDye Terminator chemistry on a 3,730× DNA analyzer (Thermo Fisher Scientific, Waltham, MA, USA). Sequences were assembled using Sequencer software V4.9 (GeneCodes, Ann Arbor, MI, USA) and compared to the CYP2D6 accessions {"type":"entrez-nucleotide","attrs":{"text":"M33388.1","term_id":"181303","term_text":"M33388.1"}}M33388.1 and {"type":"entrez-nucleotide","attrs":{"text":"AY545216","term_id":"45024927","term_text":"AY545216"}}AY545216.³⁹

To determine the haplotypes of two novel subvariants of known CYP2D6 haplotypes in subject CMH396, allele-specific XL-PCR was performed with primer −740C>T to generate a 5.5-kb XL-PCR product from the CYP2D6*1 variant as described.⁴⁶

WGS was performed as previously described.¹⁹ Briefly, 1,000 ng of DNA was sheared to an average size of 350 nt using a Covaris S2 Biodisruptor, end repaired, A-tailed and adaptor ligated using Illumina TruSeq PCR free according to manufacuter’s protocol (Illumina Inc., San Diego, CA, USA). Quantitation was performed using real-time PCR. Samples for WGS were sequenced on HiSeq 2,500 instruments (Illumina, San Diego, CA, USA) on rapid run or high-throughput mode to a read depth of ~30× from ~100 GB of total data with 2×100 nt or 2×125 nt reads. Samples were aligned and variants called with GSNAP and the Genome Analysis Tool kit (GATK), respectively,^47–49 relative to the GRCh37 CYP2D6*2 reference, yielding 5.1 million variants per genome as a variant call format (.vcf) file (Supplementary Table 2). Variants were called using methods previously described,¹⁹ briefly positions were downsampled to 750 reads using bases with sequence quality ≥20, mapping quality ≥17 and with a minimum phred-scaled confidence score of 20.0. Subsequently, variants were compared with the standard CYP2D6*1 reference ({"type":"entrez-nucleotide","attrs":{"text":"AY545216","term_id":"45024927","term_text":"AY545216"}}AY545216) allele.