How Many Cocrystals Are We Missing? Assessing Two Crystal Engineering Approaches to Pharmaceutical Cocrystal Screening

Drug development may include extensive screening for crystalline forms of active pharmaceutical ingredients. Crystal engineering aims to apply supramolecular knowledge to simplify such a task. The failure of such a strategy may result in overlooking potentially interesting compounds. Here, the advantages of such a knowledge-based approach is compared to a systematic crystallization screening for pharmaceutical cocrystals. This work indicates that a screening simply based on known synthons and their relative frequency as reported in the database is as effective as a random screening exercise, potentially missing 25% of the successful cocrystallization. Readily available computational methods perform better, enabling the identification of all of the observed cocrystals with a reduction of 24% of the experimental attempts.