Predictive Modeling of Tacrolimus Dose Requirement Based on High-Throughput Genetic Screening

• DOI: 10.1111/ajt.14040
• Publication Date: 2016-09-06T08:17:46Z
• Authors: C. Damon; M. Luck; L. Toullec; I. Etienne; M. Buchler; B. Hurault de Ligny; G. Choukroun; A. Thierry; C. Vigneau; B. Moulin; A.-E. Heng; J.-F. Subra; C. Legendre; A. Monnot; A. Yartseva; M. Bateson; P. Laurent-Puig; D. Anglicheau; P. Beaune; M.A. Loriot; E. Thervet; N. Pallet
• Keywords: False Discovery Rate

Any biochemical reaction underlying drug metabolism depends on individual gene-drug interactions and groups of genes interacting together. Based a high-throughput genetic approach, we sought to identify set covariant single-nucleotide polymorphisms predictive interindividual tacrolimus (Tac) dose requirement variability. Tac blood concentrations (Tac C0 ) 229 kidney transplant recipients were repeatedly monitored after transplantation over 3 mo. Given the high dimension genomic data in comparison low number observations multicolinearity among variables (gene variants), developed an original approach that integrates ensemble variable-selection strategy reinforce stability process multivariate modeling. Our models explained up 70% total variability per with maximum 44 gene variants (p-value <0.001 permutation test). These included molecular networks oxidoreductase activities multidrug-resistant ABCC8 transporter, which was found most stringent model. Finally, identified intronic variant encoding SLC28A3, as key involved metabolism, confirmed it independent validation cohort.