Computationally predicting prokaryotic protein subcellular localization, using experimentally-based criteria

计算机预测位于细菌细胞表面的蛋白质在医学和农业中具有广泛的意义，以确定新的潜在可获得的(表面暴露的)药物靶点、诊断标记和候选疫苗。这种分析还有助于识别微生物环境检测的标记，从耐热/耐冷的古生代鉴定具有工业意义的蛋白质的稳定性，一般的基因组注释，以及蛋白质组实验的设计。此前，我领导了PSORTb的开发，它是目前细菌和古菌蛋白亚细胞定位(包括细胞表面蛋白定位)最准确的预测指标。 然而，PSORTb还需要改进，以更好地处理原核生物的多样性及其不同的细胞结构。特别是，某些医学上非常重要的细菌，如分枝杆菌，以及具有“非典型”膜结构的古生菌，预测尤其糟糕。此外，需要对暴露在细胞表面的细胞表面蛋白的成分进行更准确的预测，以用于疫苗发现或医疗或环境用途的诊断标记。 因此，我建议从文献和我自己的实验室实验中收集关于精选原核细胞表面蛋白的高质量蛋白质亚细胞定位数据。这一丰富的数据集将在开发新版本的PSORTb方面发挥关键作用，该版本将为更多种类的原核生物做出更准确的细胞表面蛋白预测，并在某些情况下预测细胞表面蛋白的拓扑结构。实验室分析，包括对构建的外膜蛋白变体的新研究，还将提供关于一种在外膜蛋白组装中起关键作用的基本细菌外膜蛋白的结构见解。生物信息学研究将产生一个强大的新的蛋白质定位预测器和相关数据库，这将对需要识别细菌和古菌细胞表面蛋白质以及蛋白质亚细胞定位的不同研究人员提供重要帮助。

Computationally predicting proteins located on the surface of bacterial cells is of wide interest in both medicine and agriculture, to identify new potentially accessible (surface-exposed) drug targets, diagnostic markers, and vaccine candidates. Such analysis also aids identification of markers for environmental detection of microbes, characterization of proteins of industrial interest for their stability from thermo/cold tolerant Archaea, general genome annotation, and design of proteomic experiments. Previously, I led the development of PSORTb, which is currently the most precise predictor of bacterial and archaeal protein subcellular localization, including cell surface protein localization. However, PSORTb needs to be improved to better handle the diversity of prokaryotes and their different cell structures. In particular, certain very medically important bacteria like Mycobacteria spp., as well as Archaea with 'atypical' membrane structures, are particularly poorly predicted. In addition, more accurate prediction of the components of cell surface proteins that are exposed on the cell surface is needed, for vaccine discovery or diagnostic markers for medical or environmental use. Therefore, I propose to gather high quality protein subcellular localization data about select prokaryotic cell surface proteins, both from the literature and my own laboratory experimentation. This rich data set will play a key role in development a new version of PSORTb that will make more accurate cell surface protein predictions for more diverse prokaryotes, and in some cases predict cell surface protein topology. The laboratory analysis, including a novel study of constructed outer membrane protein variants, will also provide structural insights regarding an essential bacterial outer membrane protein that plays a key role in outer membrane protein assembly. The bioinformatics research will generate a powerful new predictor and associated database of protein localization which will be a significant aid to diverse researchers needing to identify bacterial and archaeal cell surface proteins, and protein subcellular localization in general.