CAREER: Computing accurate free energies for solubility, solvation, and transfer

加州大学欧文分校的大卫·莫布利获得了化学部和高级网络基础设施部化学理论、模型和计算方法计划颁发的奖项的支持，该奖项旨在开发、改进和使用计算机技术在新分子产生之前预测它们的性质。目前，在许多情况下，必须制造和测试新的感兴趣的分子，以确定它们的性质，如它们的溶解性，它们可能在环境中的分布，或者它们与重要生物分子的相互作用。但制造(合成)新分子的过程可能既缓慢又昂贵，而且有了合适的计算机工具(如本项目中的那些先进工具)，可以在合成之前通过计算预测它们的性质。这可以大大节省时间和金钱，因为只有具有理想性能的化合物才能合成。该项目专注于开发和应用计算机工具来预测这些性质，特别强调三个具体领域：预测溶解度，自动化计算以预测结合相互作用，以及预测分配到不同环境中(即在了解环境分布和影响方面)。一个长期的目标是为研究人员提供经过验证的计算工具，他们可以使用这些工具来指导研究工作设计新的分子。拟议的工作开发并应用了有前途的第一性原理计算方法来预测热力学性质，如溶解度(例如，在水中溶解固体)和分子在溶液中的溶剂化热力学。这项工作建立在莫布利博士之前在该领域的方法论创新的基础上。主要的新进展包括一个基于力场的物理溶解度预测框架，它将用多个力场进行测试，以及一种进行相对自由能计算的新的自动化方法。重点放在软件和工作流程开发上，以确保开发和应用的技术很容易为其他研究人员所用。这项工作的方法和改进将通过以教程和背景为特色的合作alChemistry.org网站进行传播，并通过调查人员将通过Simtk.org、alChemistry.org和其他地方发布的软件工具传播。

David Mobley, of the University of California, Irvine is supported by an award from the Chemical Theory, Models and Computational Methods program in the Chemistry Division and the Division of Advanced Cyberinfrastructure to develop, improve and use computer techniques to predict the properties of new molecules before they are created. Currently, in many cases, new molecules of interest must be made and tested to determine their properties such as their solubility, their likely distribution into the environment, or their interactions with important biomolecules. But the process of making (synthesizing) new molecules can be slow and expensive, and with the right computer tools (such as those advanced in this project), their properties can be predicted computationally in advance of their synthesis. This can result in substantial savings of time and money, since only compounds with desirable properties can then be synthesized. This project focuses on developing and applying computer tools to predict these properties, with a particular emphasis on three specific areas: Predicting solubilities, automating calculations to predict binding interactions, and predicting partitioning into different environments (i.e. in understanding environmental distribution and effects). A long-term goal is to provide researchers with well-validated computational tools which they can use to help guide research efforts designing new molecules.The proposed work develops and applies promising first-principles computational methods to predict thermodynamic properties such as solubilities (e.g. for dissolving a solid in water) and solvation thermodynamics for molecules in solution. This work builds on Dr. Mobley's previous methodological innovations in the area. Major new developments include a framework for physical, force-field-based solubility predictions, which will be tested with multiple force fields, and a new automated approach for conducting relative free energy calculations. A substantial emphasis is placed on software and workflow development to ensure the techniques developed and applied are easily accessible to other researchers. Methods and improvements from this work will be disseminated through the collaborative alchemistry.org website, featuring tutorials and background, and through software tools the investigator will release via Simtk.org, alchemistry.org, and elsewhere.