Inference-assisted intelligent crystallography based on preliminary data

Crystal structure analysis is routinely used to determine atomically resolved molecular structures and structure-property relationships. The accumulation of reliable structural characteristics obtained by crystal structure analysis has forged a robust basis that is frequently used in molecular and materials sciences. However, experimental techniques remain hampered by time-consuming ‘blind’ measurement-analysis iterations, which are sometimes required to find appropriate crystals and experimental conditions. Herein, we present a method that uses a small preliminary data set to evaluate the to-be-observed structures and the to-be-collected data. Moreover, we demonstrate the practical utility of this method to improve the efficiency of crystal structure analysis. This method will help selecting suitable crystals and choosing favorable experimental conditions to generate results that satisfy the level of precision required for specific research objectives.