Abstract
Structure-activity relationships (SARs) are a cornerstone of drug discovery, aiming to elucidate the connection between chemical structures and their biological properties. While widely applied in medicinal chemistry and computer-aided molecular design, SAR traditionally assumes a direct connection (i.e., relations) between chemical structures and their activity. However, given the complexity of biological responses, these connections are often better described as associations rather than strict relationships. Beyond semantics, relations usually imply a deterministic or functional mapping, whereas, associations are treated with statistical tools that capture probabilistic patterns without assuming causality. Adopting an association-based perspective helps to avoid overstated claims and manage uncertainty more realistically. In this article, structure-property associations (SPAs) are proposed as a more accurate framework to capture the connection between chemical structures and their properties in the context of drug discovery. SPA is particularly emphasized in describing the associations between chemical structures and biological activity across different experimental levels, including both in vitro and in vivo assays.