Prediction of human skin permeability using artificial neural network (ANN) modeling

Aim:To develop an artificial neural network (ANN) model for predicting skinpermeability (log Kp) of new chemical entities.Methods:A large dataset of 215experimental data points was compiled from the literature.The dataset was subdi-vided into 5 subsets and 4 of them were used to train and validate an ANN model.The same 4 datasets were also used to build a multiple linear regression (MLR)model.The remaining dataset was then used to test the 2 models.Abrahamdescriptors were employed as inputs into the 2 models.Model predictions werecompared with the experimental results.In addition,the relationship betweenlog Kpand Abraham descriptors were investigated.Results:The regression re-sults of the MLR model were n=215,determination coefficient (R~2)=0.699,meansquare error (MSE)=0.243,and F=493.556.The ANN model gave improved resultswith n=215,R~2=0.832,MSE=0.136,and F= 1050.653.The ANN model suggests thatthe relationship between log Kpand Abraham descriptors is non-linear.Conclusion:The study suggests that Abraham descriptors may be used to predict skinpermeability,and the ANN model gives improved prediction of skin permeability.