Quantitative structure retention relationship modeling in liquid chromatography method for separation of candesartan cilexetil and its degradation products

• Forced degradation studies and confirmation of identity of degradation products.
• Descriptor calculation selection based on correlation with one another.
• Two Box-Behnken designs to obtain data for network training and validation.
• QSRR was established using artificial neural networks.
• Finding optimal conditions and experimental confirmation.

Artificial neural network (ANN) is a learning system based on a computation technique, which was employed for building of the quantitative structure-retention relationship (QSRR) model for candesartan cilexetil and its degradation products. Candesartan cilexetil has been exposed to forced degradation conditions and degradation products have been subsequently identified with the assistance of HPLC-MS technique. Molecular descriptors have been computed for all compounds and were optimized together with significant chromatographic parameters employing developed QSRR models. In this way, QSRR has been used in development of HPLC stability-indicating method, optimal conditions toward various outputs have been established and high prediction potential of the created QSRR models has been proved.