Research ArticlePharmacogenomics

Toward predicting CYP2D6-mediated variable drug response from CYP2D6 gene sequencing data

See allHide authors and affiliations

Science Translational Medicine  21 Jul 2021:
Vol. 13, Issue 603, eabf3637
DOI: 10.1126/scitranslmed.abf3637

You are currently viewing the editor's summary.

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution

A prescription for dosing prediction

The highly polymorphic enzyme cytochrome P450 2D6 (CYP2D6) processes about a quarter of clinically used drugs. However, its activity varies widely between individuals, affecting drug response. Useful but imperfect *-allele categories are currently used to estimate per-individual CYP2D6 activity and refine patient drug dosages as needed. Instead, van der Lee et al. trained a neural network on entire CYP2D6 gene sequences to model enzyme activity on a continuous scale. When applied to independent cohorts of patients treated with tamoxifen or venlafaxine (both CYP2D6 substrates), the model resulted in a moderate improvement in the prediction of individual drug response, laying the groundwork for more accurate CYP2D6 pharmacogenomics.

View Full Text

Stay Connected to Science Translational Medicine