Comparison of prediction models for blood brain barrier permeability and analysis of the molecular descriptors.

To develop an in silico model for predicting blood brain barrier (BBB) permeability, and evaluate whether incorporation of two new biological descriptors, high affinity P-glycoprotein substrate probability (HAPSP) and plasma protein binding ratio (PPBR), could result in a better model, four different multiple linear regression (MLR) models have been constructed and compared with each other. The optimized model demonstrated predictive ability and simplicity, not only suitable for passive but also for active transport. Moreover, the molecular descriptors used here are discussed.