蛋白质二级结构预测中的简化编码技术

Simplified Encoding of Secondary Structure Prediction

神经网络用于蛋白质二级结构预测时,通常氨基酸序列采用正交二进制编码。基于不同残基间的物理化学性质,提出了简化的编码技术,并与其他蛋白质二级结构预测的方法进行了比较。实施结果表明:这种方法更充分地利用了蛋白质一级结构的信息,有较好的效果。

Orthogonal encoding mode has often been used in protein secondary st ructure prediction in neural network. Based on the physical-chemical properties of different residues, this paper puts forward a new kind of simplified encodin g technique, and makes comparison with other encoding methods. The practicing re sults show that the effect of this method is better because of making full use o f the information of first-level structure of protein.