Prediction of protein domain structural classes is an important topic in protein science. In this paper, we proposed a new conception: structural class tendency of polypeptides (SCTP), which is based on the fact that a given amino acid fragment tends to be presented in certain type of proteins. The SCTP is obtained from an available training data set PDB40-B. When using the SCTP to predict protein structural classes by Intimate Sorting predictive method, we got the predictive accuracy (jackknife test) with 93.7%, 96.5%, and 78.6% for the testing data set PDB40-j, Chou&Maggiora and CHOU. These results indicate that the SCTP approach is quite encouraging and promising. This new conception provides an effective tool to extract valuable information from protein sequences.

中文翻译：

---

多肽结构类别趋势：预测蛋白质结构类别的新概念

蛋白质结构域的预测是蛋白质科学中的一个重要课题。在本文中，我们提出了一个新的概念：多肽的结构类别趋势（SCTP），它基于给定的氨基酸片段倾向于呈现在某些类型的蛋白质中的事实。SCTP 是从可用的训练数据集 PDB40-B 中获得的。当使用 SCTP 通过 Intimate Sorting 预测方法预测蛋白质结构类别时，我们对测试数据集 PDB40-j、Chou&Maggiora 和 CHOU 的预测准确度（折刀检验）分别为 93.7%、96.5% 和 78.6%。这些结果表明，SCTP 方法非常令人鼓舞和有希望。这一新概念提供了一种从蛋白质序列中提取有价值信息的有效工具。