State Key Laboratory of Cardiology and Medical Innovation Center, Institute for Regenerative Medicine, Department of Neurosurgery, Shanghai East Hospital, Shanghai Key Laboratory of Signaling and Disease Research, Frontier Science Center for Stem Cell Research, School of Life Sciences and Technology, Tongji University, Shanghai, 200092, China.
Interdisciplinary Research Center on Biology and Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai, 201210, China.
Nat Commun. 2024 Mar 8;15(1):2147. doi: 10.1038/s41467-024-46445-y.
The burgeoning comprehension of protein phase separation (PS) has ushered in a wealth of bioinformatics tools for the prediction of phase-separating proteins (PSPs). These tools often skew towards PSPs with a high content of intrinsically disordered regions (IDRs), thus frequently undervaluing potential PSPs without IDRs. Nonetheless, PS is not only steered by IDRs but also by the structured modular domains and interactions that aren't necessarily reflected in amino acid sequences. In this work, we introduce PSPire, a machine learning predictor that incorporates both residue-level and structure-level features for the precise prediction of PSPs. Compared to current PSP predictors, PSPire shows a notable improvement in identifying PSPs without IDRs, which underscores the crucial role of non-IDR, structure-based characteristics in multivalent interactions throughout the PS process. Additionally, our biological validation experiments substantiate the predictive capacity of PSPire, with 9 out of 11 chosen candidate PSPs confirmed to form condensates within cells.
蛋白质液-液相分离（PS）的理解不断深入，为预测具有液-液相分离潜力的蛋白质（PSP）提供了大量生物信息学工具。这些工具通常偏向于富含无序区域（IDR）的 PSP，因此经常低估了没有 IDR 的潜在 PSP。尽管如此，PS 不仅由 IDR 驱动，还由结构模块域和相互作用驱动，这些在氨基酸序列中不一定反映出来。在这项工作中，我们引入了 PSPire，这是一种机器学习预测器，它结合了残基和结构水平的特征，用于精确预测 PSP。与现有的 PSP 预测器相比，PSPire 在识别没有 IDR 的 PSP 方面有显著的提高，这突出了非 IDR、基于结构的特征在 PS 过程中的多价相互作用中的关键作用。此外，我们的生物学验证实验证实了 PSPire 的预测能力，11 个候选 PSP 中有 9 个被证实能在细胞内形成凝聚物。
