PyDescriptorC*: A Descriptor Calculation Tool for Decoding Chirality Cliffs and Revealing Hidden Patterns in Drug Discovery

Chirality plays a crucial role in the biological activity and pharmacological properties of molecules, often leading to significant differences in activity profiles, referred to as chirality cliffs or activity cliffs. To address this challenge, we introduce PyDescriptorC*, a novel Python-based script designed to calculate thousands of chirality-aware descriptors and other molecular descriptors using molecular 3D structures (total 112,194 molecular descriptors). PyDescriptorC* leverages multiprocessing, PyMOL integration, and compatibility with mol2 file formats, ensuring computational efficiency and broad applicability. The tool was rigorously tested on two structurally diverse datasets (HDAC4 and ADM17) for regression analysis, demonstrating exceptional performance in capturing chirality-driven insights and elucidating the molecular basis of chirality cliffs. These chirality descriptors provided a deeper understanding of activity patterns, significantly enhancing the predictive accuracy and interpretability of QSAR models. PyDescriptorC* has been validated across multiple operating systems, ensuring platform independence and user accessibility. This work highlights the potential of PyDescriptorC*, an extended version of PyDescriptor, to advance drug discovery and development by unlocking hidden chirality patterns and bridging the gap between structural data and predictive modelling. Its seamless integration of chirality-specific descriptors into QSAR workflows offers a powerful resource for cheminformatics, machine learning, and molecular design