這研究的主要目的預測一組蛋白質功能與蛋白質功能相關性的規則。當知道蛋白質定義的功能且此蛋白質與另一蛋白質交互作用時,可以用這一組規則來預測另一蛋白質的生物功能。相關性規則的預測是以蛋白質功能組合的方法來建置。而我們發現具有特定類型的蛋白質功能組合會傾向在一起交互作用,舉例來說,當蛋白質有轉錄的功能,此蛋白質會有很高的機會和另一個具有轉錄功能的蛋白質交互作用。 我們使用一隨機的方法模擬蛋白質的功能與功能間相關性(function-function correlation, FFC)。將兩蛋白質功能之相關性與隨機的結果作比較,得出預測蛋白質功能相關性的計分測準。 研究結果建置成網頁,供使用者查詢。此網頁提供下列功能:(i) 蛋白質功能相關性(FFC)之分數,(ii)與隨機模型比較的分數及(iii)蛋白質功能預測之服務。網頁位址如下:http://210.70.82.82/ffc。
The main purpose of this research is to derive a set of rules for protein function and protein function correlation. This set of relations is used to predict the biological functions of a protein given that it interacts with another protein having well-define functions. The calculation is based on modeling the protein-protein interaction data by the protein function combination model. It is found that certain types of protein functions tend to correlated with each other, for example, our calculations suggested that the most likely functional correlation pair is protein having transcriptional function interacts with protein has transcriptional function. Randomized version of the function combination model is performed in order to justify the function-function correlation (FFC) calculation. The correlation strength of two protein functions are compared to its randomized counterpart, hence, resulted in a scoring value to rank the relative importance of a FFC. A web based service is set up to provide the following functionality, (i) a ranking list of FFC, (ii) the relative strength of a FFC, and (iii) a prediction service for protein function. The URL address of the service is http://210.70.82.82/ffc.