Exploring the Potential of AlphaFold Distograms for Flexibility Assignment in Cryo-EM Maps

AlphaFold (AF) models are widely used in cryo-EM workflows, but their design to produce a single static structure limits the interpretation of flexible or unresolved regions in EM maps. To address this, we investigate whether AF-generated distance probability distributions (distograms) can capture conformational flexibility. As a test case, we focus on the AIFM1/AK2 complex, where molecular dynamics simulations identified a hinge-driven motion in AK2 upon AIFM1 binding. Remarkably, this conformational change is reflected in distograms from AF2.3 and AF3, although it is not captured in their final predicted structures. Interestingly, enhanced sampling and MSA perturbations failed to improve flexibility detection in earlier AF2 versions. All in all, by analyzing this key metabolic complex, we demonstrate that distograms offer a promising, structure-free approach for identifying flexible protein regions and interpreting dynamic signals in cryo-EM maps.