Quantitative Structure and Activity Relationship (QSAR) Modeling and Analysis and Design of New Compounds

Quantitative Structure–Activity Relationship (QSAR), or Hubungan Kuantitatif Struktur– Aktivitas (HKSA), is an important computational method in medicinal chemistry used to quantitatively link molecular structures with their biological activities. This method facilitates the efficient design of new compounds. The foundation of QSAR involves the use of regression models such as Multiple Linear Regression (MLR) or Partial Least Squares (PLS), as well as 2D/3D-QSAR approaches such as Comparative Molecular Field Analysis (CoMFA). The main objective is to predict molecular properties such as IC₅₀ using theoretical descriptors, including topological, geometrical, and quantum-chemical indices such as HOMO and LUMO, which are calculated through Density Functional Theory (DFT) or semi-empirical methods. In addition, physicochemical properties such as logP and pKa are calculated using computational tools such as Gaussian or RDKit. QSAR models are built using feature selection techniques and machine learning approaches, including Artificial Neural Networks (ANN). To ensure accuracy and prevent overfitting, the models are validated using metrics such as R², Q², cross- validation, and Root Mean Square Error (RMSE). Model analysis also enables the identification of descriptor contributions to molecular activity, thereby accelerating the processes of iteration, virtual screening, lead compound optimization, cost reduction, and the development of safer drugs. Nonetheless, the predictions must ultimately be confirmed through experimental validation in the laboratory.