Abstract
Intrinsically Disordered Proteins (IDPs) play fundamental roles in signaling, regulation and cell homeostasis by specifically interacting with their partners. The structural characterization of these interacting regions remains challenging and requires the integration of extensive experimental information. Here we present an approach that exploits the structural information encoded in tripeptide fragments from coil regions of high-resolution structures. Our results indicate that a simple building approach that disregards the sequence context provides a good structural representation of fully disordered regions. Conversely, the description of partially structured motifs calls for the consideration of sequence-dependent structural preferences. By using NMR Residual Dipolar Couplings and SAXS data for multiple IDPs we demonstrate that the appropriate combination of these two building strategies produces ensemble models that correctly describe the secondary structural classes and the population of partially structured regions. This study paves the way for the extension of structure prediction and protein design to disordered proteins.