“…From prior pharmacogenomic researches (Table S1) and reviews, 13,17 a total of 25 candidate genes were chosen, including ABCB1 , ABCC2 , ABCG2 , CES1 , CES1P2 , CYP1A2 , CYP2A6 , CYP2B6 , CYP2C19 , CYP2C8 , CYP2C9 , CYP2D6 , CYP2J2 , CYP3A4 , CYP3A5 , CYP4F2 , FRAS1 , SLC22A1 , SLC4A4 , SLCO1B1 , SULT1A1 , UGT1A1 , UGT1A9 , UGT2B7 and UGT2B15 . The detailed methods about genome‐wide and candidate gene association analyses were similar to those used in our GWA study on healthy participants with dabigatran 22 . Regional plots were created in LocusZoom, 28 and additional graphics were generated using R. Except otherwise specified, continuous variables were presented as mean ± standard deviation, and a two‐sided p ‐value of less than .05 was regarded as statistically significant.…”