An exact description of five-membered ring configurations. II. Applications to furanose rings in DNA and RNA, analysis of errors, and bond angle bending energy.
An exact description of five-membered ring configurations. II. Applications to furanose rings in DNA and RNA, analysis of errors, and bond angle bending energy.
复制标题
五元环构型的精确描述。
DOI：
10.1080/07391102.1993.10508698
复制
发表时间：
1993
影响因子：
4.4
通讯作者：
Day,LA
中科院分区：
文献类型：
--
作者：
Marzec,CJ;Day,LA
在论文I（在这个问题上的前一篇论文）中开发的五原子呋喃糖环的精确参数化方法已被应用于大分子结构中的环。这些数据包括A型和B型DNA中的322个环的坐标; Z型DNA中的60个环; RNA分子中的29个环;以及两种特定DNA-蛋白质复合物中的53个环。发现大分子数据集的内部坐标[q，P，S，Γ; {bj}]的特征分布与小分子数据集的相应分布非常相似，即使原子坐标中的不确定性对于较大的分子更大。为了帮助解释这些数据，给出了经验法则。分析表明，方向角的分布在很大程度上取决于键角弯曲势。这一分析结果得到了数据的充分证实，数据表明平均Γ值大致接近90°。实验观察到的每个环的Γ趋于落在理论上的Γ附近，通过最小化键角弯曲势，同时保持其他内部参数不变，来计算它。通过考虑使用完整参数化的大型数据集，可以看出这种明确无误的趋势。在输入笛卡尔坐标的不确定性的影响进行了处理，它表明，新的参数S和Γ是更敏感的误差比q和P，但在S和Γ的不确定性降低时，S是大的。计算从笛卡尔坐标到坐标[q，P，S，T; {bj}]的变换的雅可比矩阵J。可用于统计研究的雅可比矩阵有助于理解配置空间的采样、S值的分布和误差的传播。
The method developed in Paper I (the preceding paper in this issue) for the exact parameterization of the five-atom furanose ring has been applied to rings in macro- molecular structures. The data include coordinates for 322 rings in A- and B-form DNA; 60 rings in Z-form DNA; 29 rings in RNA molecules; and 53 rings in two specific DNA-protein complexes. Characteristic distributions of internal coordinates [q, P, S, Γ; {bj}] for the macromolecular data sets are found to be closely similar to the corresponding distributions for the small molecule data sets, even though the uncertainties in the atomic coordinates are larger for the larger molecules. To aid in the interpretation of such data rules of thumb are given. It is shown, analytically, that the distribution of the direction angle Γ is largely determined by the bond angle bending potential. This analytical result is fully confirmed by the data, which show that the average Γ value is roughly near 90°. The experimentally observed Γ for each ring tends to fall near the theoretical Γmincalculated for it by minimizing the bond angle bending potential while holding the other internal parameters constant. Such unmistakable trends are discernible by considering large data sets using the full parameterization. The effects of uncertainties in the input Cartesian coordinates are treated, and it is shown that the new parameters S and Γ are more sensitive to error than are q and P, but that the uncertainties in S and Γ decrease when S is large. The Jacobian J of the transformation from Cartesian coordinates to the coordinates [q, P, S, Γ; {bj}] is calculated. The Jacobian, which can be used in statistical studies, aids in understanding the sampling of configuration space, the distribution of S values, and the propagation of error.