PROTEIN INTERACTIONS: PROTEIN MICROARRAYS

本子项目是利用由NIH/NCRR资助的中心赠款提供的资源的众多研究子项目之一。子项目和研究者（PI）可能已经从另一个NIH来源获得了主要资金，因此可以在其他CRISP条目中表示。列出的机构是中心的，不一定是研究者的机构。尽管大规模蛋白质相互作用技术提供了大量有用的信息，包括酵母双杂交筛选和基于质谱的蛋白质复合物测定，但越来越明显的是，它们有两个主要缺点。它们都能检测出假阳性，而且都错过了许多生物学上相关的相互作用。假阳性和假阴性的实际频率是一些争论的主题，可能取决于技术及其应用方式，但很明显，所有方法都存在这两个问题。对大规模酵母相互作用数据集彼此之间以及与已发表数据的比较提供了一些重要的见解（von Mering et al., 2002）。首先，与已发表数据和链接功能注释的相关性表明，大量高通量数据，至少来自酵母研究，包含真正的阳性；也就是说，它并不是大部分都是垃圾。其次，数据集之间重叠的一小部分表明每种方法都容易丢失交互（接受没有数据集主要由假阳性组成）。最后，由两种或多种方法中发现的相互作用组成的重叠数据被发现包含最高的真阳性频率。综上所述，这些发现强化了这样一种观点，即需要多种高通量蛋白质相互作用技术来进行交叉验证，并确保相互作用数据的全面收集。在这里，我们建议使用蛋白质微阵列来验证来自酵母双杂交和MS-based复合物测定的蛋白质相互作用数据，并检测新的蛋白质相互作用。

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Despite the wealth of useful information coming from large scale protein interaction technologies, including yeast two-hybrid screens and MS-based protein complex determinations, it is becoming increasingly apparent that they have two major shortcomings. They all detect false positives and they all miss many biologically relevant interactions. The actual frequency of false positives and false negatives is the subject of some debate, and likely depends on the technique and how it is applied, but it is clear that all approaches have both problems. Comparisons of large scale yeast interaction data sets with each other and with published data have provided several important insights (von Mering et al., 2002). First, correlations with published data and with linked functional annotations indicate that the bulk of the high throughput data, at least from the yeast studies, contains true positives; i.e., it is not mostly junk. Second, the small fraction of overlap between datasets suggests that each method is prone to missing interactions (accepting that no dataset consists mostly of false positives). And finally, the overlapping data consisting of interactions found in two or more approaches, was found to contain the highest frequency of true positives. Combined, these findings reinforce the idea that multiple high throughput protein interactions technologies will be needed both for cross-validation, and to ensure a comprehensive collection of interaction data. Here we propose to use protein microarrays to validate protein interaction data derived from yeast two-hybrid and MS-based complex determinations, and to detect new protein interactions.