In-silico prediction of Cellular Responses to Polymeric Biomaterials from Their Molecular Descriptors

Abstract

In this work quantitative structure activity relationship (QSAR) methodology was applied for modeling and prediction of cellular response to polymers that have been designed for tissue engineering. After calculation and screening of molecular descriptors, linear and nonlinear models were developed by using multiple linear regressions (MLR) and artificial neural network (ANN) methods. The root mean square error (RMSE) of these models were RMSE_MLR=12.6 and RMSE_ANN=10.6. Robustness and reliability of the developed MLR and ANN models were evaluated by using the leave-one-out and leave many out cross-validation methods, which produces the statistics of Q²_MLR=0.74 and Q²_ANN=0.81. Moreover, the chemical applicability domains of these models were determined via leverage approach. The results of these tests indicate the suitability of developed models. Comparison of statistical parameters of MLR and ANN models indicate the suitability of non-linear over linear model. The results of this study revealed the high applicability of QSAR approach in prediction of cellular response to the polymeric biomaterials.