Artificial neural network (ANN) based modelling for D1 like and D2 like dopamine receptor affinity and selectivity.

Dopamine and its receptors play a critical role in diseases such as Parkinson's disease and schizophrenia. A problem with developing specific drugs for such diseases is that there are five subtypes of dopamine receptors that can be categorized as either D1 like or D2 like. Since the binding sites are quite similar, it is difficult to design the subtype specific agonists and antagonists required for therapy with minimal side effects. Thus, the aim of this study was to identify the molecular characteristics important to the selective binding of dopamine D1 like and D2 like receptors using quantitative structure activity relationships (QSARs). Datasets of 29 and 69 molecules capable of binding to cloned human D1 and D2 receptors were used to build QSAR models. The dissociation constants (Ki) for these molecules were taken from the literature. The optimized 3D structure of each molecule was encoded with 62 theoretical molecular descriptors. The QSAR, using hybrid neural network modeling, was built using categorical and continuous molecular descriptors as inputs, with dissociation constants (Ki) as outputs. Categorically assigned molecular descriptors improved performance in both models. Secondary amines and other nitrogen-containing moieties were shown to be important for the D1 like receptor selectivity, whereas molecular size, volume and tertiary and quaternary carbons were found to be of significant importance for the D2 like receptor selectivity.