The ultrashort X-ray pulses afforded by free-electron lasers have taken time-resolved protein crystallography into a new world full of sub-picosecond structural dynamics. As was anticipated, this has allowed exciting and unexpected observations to be made, which are, however difficult to interpret. We will discuss some of the many reasons for this difficulty, such as the low (and unknown) occupancies of the excited states under study, the presence of mixtures of states in the crystals, and sparse sampling of time courses, as well as the counterintuitive nature of protein- and, in general, molecular dynamics. As an example, we will show such an unexpected observation : unanticipated dynamics in carboxymyoglobin photolysis.
We argue that interpreting such results should only be attempted within a framework that comprises not only structural biology, but also high-level quantum mechanics and other computational means, ultrafast- and other spectroscopic techniques, and further orthogonal methods. Moreover, such interpretations must be safeguarded by appropriate investigations of coordinate errors of the structures determined, for which methods will be discussed.​

​