Electronic Structure Theory: Algorithms, Applications and Cheminformatics

导言：目前还没有一种计算方法可以可靠地处理大分子或复杂的化学过程。合理的药物设计、新材料的合理设计、催化、溶液中的反应、大分子的激发态、表面化学和蛋白质折叠都是对计算化学家非常具有挑战性的问题。这项研究主要集中在开发新的和创新的方法来进行电子结构计算。长期目标是开发可靠的方法来处理非常大的分子，如蛋白质。这些理论也应该足够简单，以提供对化学的洞察。 电子结构理论：分子密度和分子径向密度(由我的小组定义)以及它们的梯度和拉普拉斯量，将被用来研究分子中原子和键的性质(ABIM)。使用ABIM，可以计算与分子中给定的原子或键相关的许多性质。这些研究将通过研究化学反应中化学键的变化来为化学键和结构反应性提供新的见解。分子描述符也将利用ABIM的性质进行开发。这些将提供一套新颖而有价值的相似性度量，用于药物设计等领域。将创建分子中原子密度(AIM)和径向密度的数据库，并将开发使用这些AIM来生成大分子的密度或径向密度的方法。给定这些密度，我们将开发算法来使用整个大分子的密度来执行构象能量搜索。一种QM/AIM方法，将分子分成由AIM导出的密度描述的区域，并使用波函数或密度泛函理论求解较小的区域(例如，活性中心)。这种方法在模拟溶剂化过程中也是有用的。一种电子结构理论，其中不是将电子密度划分为原子贡献，而是使用变分方法从原子密度建立分子密度。可以可靠地预测关联能量的模型将基于易于计算的属性来开发。这些模型将帮助我们开发比当前方法更快、更适用于更大系统的组合方法。 研究反应机理：我们将继续研究反应机理，特别是脱胺和脱酰胺反应。这些反应包括胞嘧啶、腺嘌呤、鸟嘌呤的脱氨基和谷氨酰胺的脱氨基。将研究溶剂和催化剂的作用。在我们的研究中，我们将应用上述方法来洞察这些反应。这些反应在诱变、致癌和高氨血症(谷氨酰胺)中起重要作用。

Introduction: No computational methodologies currently exist that can deal reliably with macromolecules or complex chemical process. Rational drug-design, rational design of new materials, catalysis, reactions in solution, excited states of large molecules, surface chemistry, and protein folding are all examples of problems that are very challenging to computational chemists. The research focuses broadly on the development of novel and innovative methods for preforming electronic structure calculations. The long-term goal is to develop reliable methodologies for dealing with very large molecules, such as proteins. The theories should also be simple enough to provide insight into the chemistry. Electronic Structure Theory: The molecular density and the molecular radial density (defined by my group) along with their gradient and Laplacian, will be used to study the properties of both Atoms and Bonds In Molecules (ABIM). With ABIM it is possible to compute a number of properties associated with a given atom or bond in a molecule. Such studies will provide new insight into chemical bonding and structure-reactivity by studying the change in chemical bonds during a chemical reaction. Molecular descriptors will also be developed using the ABIM properties. These will provide a novel and valuable set of similarity measures for use in areas such as drug design. A database of Atoms-In-Molecules (AIM) densities and radial densities will be created and methods will be developed to use these AIM to generate the density, or radial density, of macromolecules. Given these densities, we will develop algorithms to perform conformational energy searches using the density of the entire macromolecule. A QM/AIM approach, in which the molecule is divided into a region described by AIM derived densities and a smaller region (e.g., the active site) is solved using wavefunction or density functional theory. Such an approach may also be useful in modelling solvation. An electronic structure theory in which, instead of partitioning the electronic density into atomic contributions, the molecular density will be built from atomic densities using a variational approach. Models which can reliably predict the correlation energy will be developed based on properties that can be easily calculated. These models will help us develop composite methods that are fast and applicable to much larger systems then with the current methods. Investigating Reaction Mechanisms: We will continue investigating reaction mechanisms, especially deamination and deamidation reactions. These include the deamination of cytosine, adenine, guanine and the deamidation of glutamine. The role of solvent and catalysts will be investigated. In our studies we will apply the methodologies described above to provide insight into these reactions. These reactions are important in mutagenesis, carcinogenesis and hyperammonemia (glutamine).