| Summary: | 碩士 === 國立交通大學 === 分子醫學與生物工程研究所 === 105 === Antioxidant protein (AOP) is considered as biological endogenous antioxidant defense mechanisms, tends to donate electrons to scavenge Reactive oxygen species (ROS). The issue in the past has been widely disused which mostly focus on one topic of AOP, but rarely on overall. Therefore, this study proposes a novel methodological approach called SCMAOP to estimating the propensity scores of 400 dipeptides and 20 amino acids in order to design prediction methods and characterize AOPs based on a scoring card method (SCM). Moreover, using SCMAOP to finding the functional recognition pattern on antioxidant protein. The SCMAOP method for predicting AOPs achieves a test accuracy of 86.17%. A dataset consisting of known AOPs corresponding its ROS substrate both H2O2 and O22- were collected from PeroxiBase and SwissProt database. Additionally, informative physicochemical properties of 20 amino acids are identified using the estimated propensity scores to characterize AOPs as follows: 1) AOP have high propensity derived from alpha helix designed sequences. 2) The structure of AOP have a lot of beta-turn structure which may as catalytic environments. 3) The specific functional structure of AOP may located from beta-strand connecting loop-helix conformation which contain beta-turn structure. Further, we develop a tool called IDAOP based on the specific functional structure to verify our finding. The analysis results reveals that the specific secondary structure is important to antioxidant protein, beta-turn structure is the main impact factor, identifying 131 putative AOP from putative dataset after IDAOP method. According to the published studied that four putative AOP from top 50 rankings may have similar antioxidant function. Therefore, we believe that combine SCMAOP and IDAOP can help biologists to finding novel antioxidant protein.
|
|---|
