Misra G P, Wong C F
Department of Physiology and Biophysics, Mount Sinai School of Medicine, New York 10029-6574, USA.
Proteins. 1997 Jul;28(3):344-59.
A novel helix-coil transition theory has been developed. This new theory contains more types of interactions than similar theories developed earlier. The parameters of the models were obtained from a database of 351 nonhomologous proteins. No manual adjustment of the parameters was performed. The interaction parameters obtained in this manner were found to be physically meaningful, consistent with current understanding of helix stabilizing/destabilizing interactions. Novel insights into helix stabilizing/destabilizing interactions have also emerged from this analysis. The theory developed here worked well in sorting out helical residues from amino acid sequences. If the theory was forced to make prediction on every residue of a given amino acid sequence, its performance was the best among ten other secondary structural prediction algorithms in distinguishing helical residues from nonhelical ones. The theory worked even better if one only required it to make prediction on residues that were "predictable" (identifiable by the theory); > 90% predictive reliability could be achieved. The helical residues or segments identified by the helix-coil transition theory can be used as secondary structural contraints to speed up the prediction of the three-dimensional structure of a protein by reducing the dimension of a computational protein folding problem. Possible further improvements of this helix-coil transition theory are also discussed.
一种新的螺旋-卷曲转变理论已经被提出。这个新理论包含比早期提出的类似理论更多类型的相互作用。模型的参数是从一个包含351个非同源蛋白质的数据库中获得的。没有对参数进行人工调整。以这种方式获得的相互作用参数被发现具有物理意义，与目前对螺旋稳定/不稳定相互作用的理解一致。从这个分析中也出现了对螺旋稳定/不稳定相互作用的新见解。这里提出的理论在从氨基酸序列中筛选出螺旋残基方面表现良好。如果该理论被迫对给定氨基酸序列的每个残基进行预测，在区分螺旋残基和非螺旋残基方面，其性能在其他十种二级结构预测算法中是最好的。如果只要求该理论对“可预测”（由该理论识别）的残基进行预测，该理论的效果会更好；可以实现>90%的预测可靠性。由螺旋-卷曲转变理论识别出的螺旋残基或片段可以用作二级结构约束，通过降低计算蛋白质折叠问题的维度来加速蛋白质三维结构的预测。还讨论了这种螺旋-卷曲转变理论可能的进一步改进。
