Cornell Michael, Paton Norman W, Oliver Stephen G
Department of Computer Science, University of Manchester, Kilburn Building, Oxford Road, Manchester M13 9PL, UK.
Comp Funct Genomics. 2004;5(5):382-402. doi: 10.1002/cfg.412.
Global studies of protein-protein interactions are crucial to both elucidating gene function and producing an integrated view of the workings of living cells. High-throughput studies of the yeast interactome have been performed using both genetic and biochemical screens. Despite their size, the overlap between these experimental datasets is very limited. This could be due to each approach sampling only a small fraction of the total interactome. Alternatively, a large proportion of the data from these screens may represent false-positive interactions. We have used the Genome Information Management System (GIMS) to integrate interactome datasets with transcriptome and protein annotation data and have found significant evidence that the proportion of false-positive results is high. Not all high-throughput datasets are similarly contaminated, and the tandem affinity purification (TAP) approach appears to yield a high proportion of reliable interactions for which corroborating evidence is available. From our integrative analyses, we have generated a set of verified interactome data for yeast.
对蛋白质-蛋白质相互作用进行全球范围的研究对于阐明基因功能以及形成对活细胞运作机制的整体认识都至关重要。已经利用遗传筛选和生化筛选对酵母互作组进行了高通量研究。尽管这些实验数据集规模庞大，但它们之间的重叠非常有限。这可能是由于每种方法仅对总互作组的一小部分进行了抽样。或者，这些筛选产生的大部分数据可能代表假阳性相互作用。我们利用基因组信息管理系统（GIMS）将互作组数据集与转录组和蛋白质注释数据进行整合，发现有重要证据表明假阳性结果的比例很高。并非所有高通量数据集都受到类似程度的污染，串联亲和纯化（TAP）方法似乎产生了很大比例的可靠相互作用，且有确证依据。通过我们的综合分析，我们生成了一组经过验证的酵母互作组数据。
