Efficient double hybrid density functional theory algorithms for conformational a

描述（由申请人提供）：双杂化（DH）密度泛函是模拟生物分子的量子力学方法中最有前途的新发展之一。据报道，DH功能比上一代功能产生更准确的分子间相互作用能。因此，它们即将被广泛使用。然而，除了对用户缺乏熟悉之外，仍然存在阻碍DH功能广泛使用的实际障碍。该SBIR提案旨在通过直接解决已知的关键限制，拥抱DH方法提供的更高准确性的机会。这些限制和我们建议的补救措施如下。(i) DH泛函在计算上比传统泛函更昂贵，因为成本是分子大小的5次幂，而不是3次幂。我们将实现一种新的三阶方法来评估DH能量，而不需要显式的双电子积分。计算节省的生物物理有用的准确性将评估构象相对能量。（ii） DH泛函需要更大的原子轨道基集来产生收敛的相对能量，但加倍基集需要16倍的计算。通过F12技术，我们将显著改善这种缓慢的收敛，通过增加基础，明确地包括电子间坐标，通过波函数方法证明是成功的F12技术。这是第一次使用F12方法来评估DH能量，将对其进行基准测试，以评估在不牺牲构象能和分子间相互作用精度的情况下可能减少的基集大小。

DESCRIPTION (provided by applicant): Double hybrid (DH) density functionals are one of the most promising new developments in quantum mechanical methods for simulating biomolecules. DH functionals have been reported to yield much more accurate intermolecular interaction energies than previous generation functionals. They are therefore on the threshold of becoming widely used. However, there are still practical obstacles blocking widespread use of DH functionals, apart from lack of familiarity to users. This SBIR proposal aims to embrace the opportunity of greater accuracy offered by the DH approach by directly addressing the key known limitations. These limitations and our proposed remedies are as follows. (i) DH functionals are more computationally expensive to evaluate than conventional functionals because the cost scales as the 5th power, rather than the 3rd power of molecular size. We will implement and benchmark a novel 3rd order method for evaluating the DH energy without explicit two-electron integrals. The computational savings for biophysically useful accuracy will be assessed on conformational relative energies. (ii) DH functionals require larger atomic orbital basis sets to yield converged relative energies, but doubling the basis requires 16 times more computation. We will significantly ameliorate this slow convergence by augmenting the basis to explicitly include interelectronic coordinates, via F12 techniques that have proven successful for wave function methods. This first use of F12 methods to evaluate the DH energy will be benchmarked to assess the reduction in basis set size that is possible without sacrificing accuracy for conformational energies and intermolecular interactions. PUBLIC HEALTH RELEVANCE: to implement a new DFT method in a computationally efficient manner. DFT is at the core of molecular modeling and is applied widely in biological research/development and in drug discovery. The improved DFT will significantly increase researchers' quality of work and extend the application scope of DFT.