QBioFusion-QSAR: Morgan-Anchored Quantum Multiple Kernel Learning for Small-Data Ligand Classification

QBioFusion-QSAR is a novel framework that integrates quantum multiple kernel learning (QMKL) with Morgan/Tanimoto fingerprints for small-data ligand classification. It combines classical and quantum kernels, achieving an accuracy improvement from 0.815 to 0.833 and an increase in Matthews correlation coefficient (MCC) from 0.613 to 0.645 on the PsychLight-A benchmark. This framework is significant for practitioners as it provides an auditable method to enhance predictive modeling in challenging QSAR scenarios, particularly in cases with activity cliffs and limited data.