Assessment of in silico models for fraction of unbound drug in human liver microsomes

Areas covered in the review: This review highlights the in silico modeling techniques for fu(mic) including physicochemical properties-based modeling, pharmacophore feature-based classification modeling and more complex statistical method-based modeling. The application of these modeling techniques to the understanding of the structure-binding relationships is also discussed. What the reader will gain: The reader will gain an understanding of the modeling techniques used for prediction of binding to human liver microsomes (fu(mic)). The reader will also understand the molecular structure-microsomal protein binding relationships. In all of these models, lipophilicity is the most important molecular property underlying the structure-binding relationship. Other molecular properties such as charge type (positive vs negative) and hydrogen bonding are also important factors for microsomal binding. Take home message: The predictive accuracy of fu(mic) models in the high lipophilicity and tight microsomal binding ranges still needs to be further improved. However, in silico models are still valuable tools to aid chemical library design and prioritize multiple chemical series, which could provide efficiency and decrease knowledge cycle times in drug discovery.