[[alternative]]藉辨認模體來預測水解酶蛋白質之功能

[[abstract]]蛋白質在生物體中扮演重要的角色，蛋白質有多種功能例如：運輸、運動、酵素、免疫、支持、保護等功能，而自從人類基因體計畫的完成，蛋白質功能的發現成為現在重要的工作之一，而蛋白質功能的發現，最正確的方法是屬於使用生物實驗所得到的結果，但是面對龐大數量的蛋白質，使用生物實驗必須消耗大量的金錢與時間，所以配合電腦預測蛋白質功能也成為重要方法之ㄧ，這些方法基本上分成三類，第一種是以序列為基準，第二種是以結構資訊為基準，最後是混合以上兩者，現在有龐大的蛋白質序列資料，大約還有三分之ㄧ沒有發現功能，但是有結構資訊的蛋白質又過少，所以本論文重點放在序列為主的蛋白質功能預測，藉著辨認出蛋白質的模體，判斷出蛋白質的功能，對於水解酶蛋白質之功能的辨識正確率已經達到99%，未來將進行對其他種類的蛋白質進行預測。Protein plays an important role in organisms. Protein has many kinds of functions, for example: transportation, sports, enzymes, immunization, support, protection, etc. After the Human Genome Project had been completed, the discovery of protein function becomes an important work. The method that uses biological experiments can achieve the most accurate results in finding the function of a protein. But lots of money and time are needed to find the functions of so many proteins. Therefore, predicting the function of a protein by computer is an important research issue. The computer methods in predicting protein function can be divided into three classes: methods based on sequences, methods based on structures, and methods that combine the previous two. In this thesis, a method based only on sequences to predict the protein function was proposed. This method predicts functions by recognizing the motifs containied in the sequence. In order to test the proposed method, an experiment has been conducted to predict the seven subfunctions of hydrolase proteins. The accuracy achieved is above 99%