基于HNP三态模型及相对熵方法的蛋白质折叠研究

Protein Folding Study Based on The HNP Model and The Relative Entropy Approach

把20种氨基酸简化为3类:疏水氨基酸(hydrophobic,H)、亲水氨基酸(hydrophilic,P)及中性氨基酸(neutral,N),每个氨基酸简化为一个点,用其C!原子来代替.采用非格点模型,以相对熵作为优化函数,进行蛋白质三维结构预测.为了与基于相对熵方法的蛋白质设计工作进行统一,采用了新的接触强度函数.选用蛋白质数据库中的天然蛋白质作为测试靶蛋白,结果表明,采用该模型和方法取得了较好的结果,预测结构相对于天然结构的均方根偏差在0.30～0.70nm之间.该工作为基于相对熵及HNP模型的蛋白质设计研究打下了基础.

Twenty kinds of amino acids are simplified into 3 types： hydrophobic amino acids （H）, hydrophilic amino acids （P） and neutral amino acids （N）. Each residue is reduced to a head which locates in the position of the Cα atom. The off-lattice model is adopted and the relative entropy is used as a minimization function to predict the tertiary structure of a protein. A new contact intensity function is given to consist with protein design research based on the relative entropy. Testing on several real proteins from Protein Data Bank （PDB） shows the good results obtained with the model and method. The root mean square deviations （RMSD） of the predicted structures relative to the native structures range from 0.30 to 0.70 nm. A foundation for studying protein design using the HNP model and the relative entropy was made.