截至目前為止，屬於高產量的蛋白質交互作用每年以指數方式成長，由於有充足的資料讓生物學家可以很簡單去建構可能性蛋白質交互作用網路，並且從中觀察蛋白質間的功能相關性和功能相似的模組化結構，但是在過程當中卻有一個非常大的問題，那就是網路往往複雜到無法用人的肉眼可以清楚分析，所以在本論文中，我們提出一個自動從蛋白質交互作用網路中系統化萃取功能模組的流程。先利用蛋白質交互作用網路的結構分析從網路中辨識出密集區域，接著針對每個密集區域，引入生物註解資訊利用最小成本擴張樹聚類演算法找出功能模組。從實驗結果當中，發現利用本論文系統所萃取出來的功能模組確實具有功能一致性特色，並且在最後做了一個真實資料集實驗，藉由整合模組化網路和表達序列標籤資料，找出標記在模組化網路上的表現樣本，因而能夠發現不同患者間表現樣本的差異性，相信利用本系統所產生的模組化網路將可以幫助生物學家利用視覺化呈現的資訊來進行更深入的研究。

Protein-protein interaction data produced by high throughput techniques grows exponentially in recent years. Biologists utilize it to build protein-protein interaction network and observe functional relationships between proteins and functional modular property. One of the major problems is that protein-protein interaction network is usually too complex to be analyzed directly. In this study, we present a systematic procedure for automated functional modules prediction. First, the dense regions can be identified from protein-protein interaction network using network structure analysis. Second, for each dense region, a MST-based clustering algorithm which integrated biological annotation is used to generate functional modules. According to the experimental results, we have discovered that functional modules extracted from our system is actually functional coherence. And we also applied our methods to a real case study. Integrating modular networks with EST data, we found the difference among different types of patients. We believe that the proposed modular networks and visualization presentation are useful for biologists in protein-protein interaction researches.

[1] Luscombe, N.M. et al. What is Bioinformatics? A proposed definition and overview of the field. Method Inform Med 2001;40:346-58.
[2] Chen, Y. and Xu, D. Computational analyses of high-throughput protein-protein interaction data. Bentham Science Publishers 2003,1389-2037/03.
[3] Auerbach, D. et al. The post-genomic era of interactive proteomics:Facts and perspectives. Proteomics 2002,1615-9853/02/0606-611.
[4] Burn, C. et al. Clustering proteins from interaction networks for the prediction of cellular functions. BMC Bioinformatic 2004, 5:95.
[5] Nabieva, E. et al. Whole-proteome prediction of protein function via graph-theoretic analysis of interaction maps. Bioinformatics 2005, Volume 21, Suppl 1.
[6] Hahn, A. et al. Confirmation of human protein interaction data by human expression data. BMC Bioinformatics 2005, 6:112.
[7] Grigoriev, A. A relationship between gene expression and protein interactions on the proteome scale: analysis of the bacteriophage T7 and the yeast Saccharomyces cerevisiae. Nucleic Acids 2001,Vol. 29,No. 17.
[8] Kemmeren, P. et al. Predicting gene function through systematic analysis and quality assessment of high-throughput data. Bioinformatics 2005,Volume 21, No 8.
[9] Edwards, A.M. et al. Bridging structural biology and genomics :assessing protein interaction data with known complexs. TRENDS in Genetics 2005,Vol. 18,No. 10.
[10] Wu, H. Prediction of functional modules based on comparative genome analysis and Gene Ontology application. Nucleic Acids 2005,Vol. 33,No 9.
[11] Gagneur, M. Modular decomposition of protein-protein interaction networks. Genome Biology 2004,5:R57.
[12] Bader, G.D. An automated method for finding molecular complex in large protein interaction networks. BMC Bioinformatics 2003, 4:2
[13] King, A.D. Protein complex prediction via cost-based clustering. Bioinformatics 2004, Volume 20, No 17
[14] Lee, H.K. et al. Coexpression analysis of human genes across many Microarray data sets. Genome Research 2004, 14:1085-1094.
[15] Tornow, S. et al. Functional modules by relating protein interaction networks and gene expression. Nucleic Acids Research 2003,Vol. 31,No. 21
[16] Xu, Y. el at. Clustering gene expression data using a graph-theoretic approach: an application of minimum spanning trees. Bioinformatics 2002, Volume 18, No 4.
[17] Chin, C.L. and Chiang, J.H. GeneLibrarian: An integrated gene Ontology clustering and biomedical text mining system. 國立成功大學碩士論文，2005
[18] W, C.W. and Chiang, J.H. Biomodular - an integrated system for functional module discovery from protein-protein interaction networks. 國立成功大學碩士論文，2005
[19] IIR Lab, GeneLibrarian http://140.116.247.50/GeneGoogling/
[20] 王惠鈞，後基因體時代之生物技術，http://bbsc.imb.sinica.edu.tw/biotech/05_07.pdf