新近的發展在對蛋白質的穩定的研究大多是根據在一些類似結構之間的蛋白質間的比較，並顯示增加樊得瓦爾相互作用力、氫鍵、鹽橋、或者偶極偶極相互作用對增加耐溫性有潛在的幫助。蛋白質Motif代表蛋白質群組內演化保存下來的區域，一般相信這些被保留的區域跟蛋白質的穩定和功能有關。 在一些類似結構之間的比較的方法或者部份物種的同源蛋白質比對以外，需要一個完整對當今的已知的蛋白質結構和序列上的耐溫性的分析。因為耐溫性分析需要蛋白質熱穩定的溫度資料，一個擁有蛋白質熱穩定資料的生物資料庫更是不可會缺的。在這篇論文裡,我們提出一個預測資訊系統,能對原核生物的結構蛋白質與蛋白質序列的熱穩定性做分析與預測。系統中包含一個原核生物最適生長溫度資料庫，命名為PGTdb。 對於已知結構的蛋白質，系統能對不同胺基酸所形成不同強度的離子對的數量做分析並且使用Bayesian classifier預測此蛋白質結構的熱穩定性。此預測系統對超耐熱蛋白質的預測準確度具有特別高的準確度。對只有序列沒有結構的蛋白質，系統在Pfam蛋白質群組之間做比較並找出有鑑別力的motif 區域。 系統成功找出一些被保留在部分蛋白質家族中具有鑑別力的溫度相關motif。研究案例顯示有鑑別力的motif與蛋白質的thermostability有關。最適的生長溫度資訊資料庫PGTdb在微生物的培養過程中非常有用，並且與很多生化物質的生產密切相關。 另外，在相應低溫和耐高溫蛋白質群組之間的比較提供與熱穩定性有關的關鍵生物化學的洞察力並且能用來測試蛋白質群組中個別結構上差異與演化的穩定性。 Recent developments in research on the stability of proteins show that thermophilic proteins generally have increased numbers of van der Waals interactions, hydrogen bonds, salt bridges, or dipole-dipole interaction potentially contribute to the thermostability of proteins, according to comparisons between some homologous structures. Protein motifs represent highly conserved regions within protein families and are generally accepted to describe critical regions required for protein stability and function. Beyond methods of comparisons between some homologous structures or limited number of genome, a complete analysis of the thermostability on current known protein structure and sequences is need. Since the thermostability analysis requires thermo-stable information of proteins, a biological database of thermo-stable information is crucial. In this Dissertation, we propose a predictive system which is capable of predicting thermostability of protein in prokaryotes and detecting discriminative motif occurrence on proteins of specific temperature, i.e., mesophilic or thermophilic. The system contains a database containing optimal growth temperatures of prokaryotic, namely PGTdb. For protein structure, the system can take input of ion-pair feature and predict the thermostability using Bayesian classifier base on ion-pair of known thermostability proteins. The prediction achieves high precision especially on hyperthermophilic protein structures. For protein sequences, the system identifies and compares corresponding mesophilic and thermophilic sequence motifs between Pfam protein families and conserved orthologous groups. The system successfully identifies some motif which only conserved on mesophilic or thermophilic sub-family but is not conserved on their counterpart sub-family. The case study shows that the discriminative motifs are related to thermostability of proteins. The optimal growth temperature information in PGTdb is very useful in cultivation of microbes, which is closely related to production of many biomaterials. Additionally, the comparisons between corresponding mesophilic and thermophilic protein families provide key biochemical insights related to thermostability and can be used to test the evolutionary robustness of individual structural comparisons of protein families.
