Genome-wide classification and identification of transmembrane protein structure and function

我们首先从文献中收集实验表征的跨膜拓扑模型，并将它们构建成数据库（TMBDB）。通过使用这一点，我们评估了几种已发表的跨膜拓扑预测方法的性能，我们开发了一种共识预测方法（ConPred），它是一些所提出的方法的组合。通过将ConPred应用于来自99个完整基因组（88个原核和12个真核）的TM蛋白质组，我们获得了52，686个具有可靠TM拓扑结构的TM蛋白质序列，并详细分析了它们的TM拓扑结构与功能的关系。结果表明，TM蛋白的功能与其TM拓扑结构密切相关，每个功能组的pfTM蛋白都有其特定的TM拓扑结构，表现为二元拓扑结构。应用二元拓扑模式法对GPCR基因进行了基因组水平的功能分类和鉴定，发现了许多新的GPCR基因。我们还研究了TM拓扑结构的进化内部基因复制，通过内部序列的相似性比较在一个单一的序列。

We first collected experimentally-characterized transmembrane topology models from literatures and constructed them into a database(TMBDB). By using this, we evaluated the performances of several published transmembrane topology prediction methods, and we developed a consensus prediction method(ConPred) which is a combination of some of the proposed methods. By applying ConPred to TM proteomes from 99 completed genomes(88 prokaryotic and 12 eukaryotic), we obtained 52,686 TM protein sequences with reliable TM topologies and analysed the TM topologies of them in detail with respect to their relation to the function. From this analysis, it was revealed that the function of TM protein is closely related to its TM topology, and each functional group pfTM proteins has its own specific TM topology pattern expressed as a binary topology patter. The genome-scale functional classification and identification of GPCRs were also earned out, by applying the binary topology pattern method, and many novel GPCR genes were found by our study. We also investigated die TM topology evolution by internal gene duplication, through the internal sequence similarity comparison within a single sequences.