| Summary: | 碩士 === 國立臺灣大學 === 資訊工程學研究所 === 92 === Motivation. Disulfide bonds play an important role in protein folding. The exact prediction of disulfide connectivity can reduce the search space in protein structure prediction. Therefore, the exact prediction of the disulfide connectivity may help the 3D structure prediction.
Result. In this paper, we proposed a simple rule to define a disulfide connectivity pattern. “The same disulfide connectivity patterns have the same distance between two cysteines in protein sequences.” We used this rule to create a disulfide profile, and then used a minimum distance scoring function to predict disulfide connectivity. We reported the experimental results in two test sets. The first test set is to compare our method with other algorithms. The second test set is to test the performance for unknown protein. In the first experiment, the value of Qp is equal to 0.49 for non-redundant proteins in test set with less than 30% sequence identity. This result is better than the other algorithms (Qp=0.44). In the second experiment, the value of Qp is equal to 0.53 for non-redundant proteins in test set with less than 30% sequence identity. Therefore, we believe that using our disulfide profile, we can achieve high accuracy in predicting unknown protein. The method proposed here is relatively simple and can generate more accurate results than conventional methods. It may also be combined with other algorithms for further improvements in disulfide connectivity prediction.
|
|---|
