Analysis and prediction of N-terminal protein sorting signals based on proteogenomics data

细胞分选信号的分析是蛋白质功能预测的基石，但我们对这些重要功能基序的理解受到缺乏实验数据的限制，并且除了少数研究充分的例子之外，即使是最好的预测方法的准确性基本上也是未知的。这些方法是在实验确认的分选信号的非常小的样本上训练的，并且在很大程度上依赖于广义信号基序，忽略了非标准情况。该项目的总体目标是扩大我们对可切割的N-末端信号的了解，包括各种类型的信号肽，线粒体靶向信号和叶绿体转运肽。基于来自广泛生物体的大量蛋白质组学数据，将实现信号肽研究的飞跃，确定、分类和分析数万个新的信号序列。这项工作将允许更好地理解在不同的细胞环境中施加在信号序列上的序列要求，以获得对其结构和功能的更完整的知识，并发现迄今为止难以捉摸的靶向序列的新原型。分析突变谱、裂解位点位置的进化保守性以及整个信号沿着进化树的增益和损失，将阐明哪些因素决定输出途径特异性、分泌效率和疾病倾向。最后，将开发用于细胞分选信号的新颖预测算法。

Analysis of cellular sorting signals is a cornerstone of protein function prediction, but our understand-ing of these important functional motifs is limited by scarce experimental data, and the accuracy of even the best available prediction methods is essentially unknown for all but a handful of well studied examples. These methods are trained on a very small sample of experimentally confirmed sorting signals and largely rely on generalized signal motifs, ignoring non-standard cases. The overall goal of the project is to expand our knowledge about cleavable N-terminal signals, including various types of signal peptides, mitochondrial targeting signals and chloroplast transit peptides. Based on massive proteogenomics data from a broad range of organisms a quantum leap in signal peptide research will be achieved, with tens of thousands of new signal sequences determined, classified, and analyzed. This work will allow to better understand sequence requirements imposed on signal sequences in different cellular contexts, to gain a more complete knowledge of their structure and function, and to discover novel archetypes of targeting sequences that have been elusive so far. Analysis of mutation spectra, evolutionary conservation of cleavage site positions as well as of gain and loss of entire signals along the evolutionary tree will shed light on what factors determine export pathway specificity, secretion efficiency, and disease propensity. Finally, novel prediction algorithms for cellular sorting signals will be developed.