QSAR without borders

摘要: Prediction of chemical bioactivity and physical properties has been one the most important applications statistical more recently, machine learning artificial intelligence methods in sciences. This field research, broadly known as quantitative structure-activity relationships (QSAR) modeling, developed many algorithms found a broad range organic medicinal chemistry past 55+ years. Perspective summarizes recent technological advances QSAR modeling but it also highlights applicability algorithms, methods, validation practices to wide research areas outside traditional boundaries including synthesis planning, nanotechnology, materials science, biomaterials, clinical informatics. As modern generate rapidly increasing amounts data, knowledge robust data-driven modelling professed within can become essential for scientists working both research. We hope that this contribution highlighting generalizable components will serve address challenge.