Structure-Activity Relationship (SAR) analysis aims to find relationships between chemical structure and biological activity (e.g. solubility, toxicity) of molecules.
SAR are derived from experimental data and their applicability can be broadened by assuming that similar molecular structures have similar activity.
Existing SAR databases can be used to predict the activity of similar molecules, which requires an efficient method to compare the similarity of molecules.
Molecules are equal if their molecular graphs are isomorphic and the labels of the atoms and bonds are equal.
PASQAL has developed a unique quantum kernel method to study graph similarity and has tested this method on the SAR type problem of toxicity screening.
Such graph kernel methods, implemented through the analog mode of PASQAL’s quantum processors, are a prime candidate for quantum advantage in the very near-term.