<scp>DR</scp> and <scp>SPIT</scp>: Statistical approaches for identifying transient structure in intrinsically disordered proteins via <scp>NMR</scp> chemical shifts

Intrinsically disordered proteins (IDPs) play key roles in various biological processes; they are associated with liquid-liquid phase separation and are targets in disorder-based drug design. Efforts to identify their structural propensities-that can be linked to molecular recognition, malfunction, targeting-still lead to ambiguous results. Secondary structure is routinely assessed by NMR spectroscopy by calculating the secondary chemical shifts (SCSs). Focusing on a given environment in the polypeptide backbone, SCSs highlight the deviation from the "random coil" state. However, the analysis is dependent on which of the numerous random coil chemical shift (RCCS) predictors is applied in the calculations, resulting in an especially pronounced ambiguity for IDPs. To overcome this, we introduce two novel statistical tools that enable the sound identification of structural propensities. We propose the chemical shift discordance ratio (DR) for prefiltering RCCS predictors based on self-consistency. Further on, we introduce the Structural Propensity Identification by t-statistics (SPIT) approach for extracting maximum information from SCS data by using multiple RCCS predictors simultaneously. This way SCS patterns indicating structural propensities can be clearly distinguished from the "noise". The applicability of these methods is demonstrated for four proteins of varying degrees of disorder. Ubiquitin and α-synuclein are used as respective benchmarks for a globular and a disordered protein, while two proline-rich IDPs are included as especially challenging molecules in secondary structure analysis.