Molecular Simulation of oil-water partition coefficients of pharma compounds

制药行业花费了大量的时间和精力（Herper, 2017）来开发潜在的候选药物，其中大多数最终由于缺乏溶解度，毒性，药效学问题以及一系列其他障碍而被抛弃。通过昂贵的实验，许多候选药物被发现是“死胡同”，尽管时间和精力投入到他们的发展。因此，开发硅替代品来寻找潜在的候选药物将是有益的，因为这将大大简化整个发现过程。多年来，分子建模领域有了巨大的发展，在原子水平上获得了有用的信息。然而，这些方法中的许多只提供了分子行为的快照。另一种方法是粗粒度（Ingólfsson, et al., 2014），它的计算速度比标准原子方法快。在这种方法中，没有明确的原子细节，而是以珠子的形式表示不同的原子群。更大的结构可以用CG方法建模。这种方法可以探索目前尚未探索的相空间，这是一个令人兴奋的领域。本博士项目旨在开发一种评估活性药物化合物（api）油水分配系数的方案。这是一项重要的研究，因为辛醇-水分配模型仍然是初步了解原料药溶解度的基础。化合物在每个相中的平衡浓度有助于确定化合物的亲脂性或亲水性（Bannan, et al., 2016）。水代表人体内的水介质，辛醇代表生物膜。因此，该模型是决定API在体内的溶解度及其在生物膜上的转运的替代品。然而，目前这是通过昂贵的实验室实验来模拟的。这个项目旨在创建一个合适的计算模型来代表这个实验。将使用粗粒度模型，其中API分子被外力通过水-油界面“拉动”。在此过程中，将监测能量势垒，以更深入地了解传输过程。监测能源障碍也将有助于确定趋势和关系（如果有的话）。这也应该有助于评估目前使用的模型，并确定为什么它可能不能完全代表药物在体内分布的实际过程。实验模型的主要局限性之一是它不能解释非平衡条件。所创建的计算模型应该有助于深入了解流程的这一部分。这个项目的很大一部分将涉及开发一个合适和准确的模型。因此，项目的第一阶段将包括对当前使用的模型（如原子模型）进行广泛的文献回顾。需要在这些模型之间进行比较，这将有助于在本项目中使用粗粒度模型。在模型可以完全建立和模拟之前，还将有其他关键的里程碑。这些目标在目标中被分成不同的类别，将在下一节中看到

The pharmaceutical industry spends a large amount of time and effort (Herper, 2017) in developing potential drug candidates, most of which are eventually discarded due to lack of solubility, toxicity, pharmacodynamics issues amongst a range of other hurdles.Through costly experimentation, many candidate drugs are found to be a 'dead-end' in spite the time and effort put into their development. Therefore, it would be beneficial to develop in silico alternatives to find potential drug candidates as this would significantly streamline the overall discovery process. Over the years there has been a vast development in the field of molecular modelling, where useful information is obtained at the atomistic level. Many of these methods, however, only provide a snapshot of molecular behaviour.An alternative methodology, which is computationally faster that the standard atomistic methods, is coarse graining (Ingólfsson, et al., 2014). In this method rather than having explicit atomistic detail there is a representation of different groups of atoms in the form of beads. Larger structures can be modelled using CG methods. This method allows the exploration of phase space currently unexplored and is an exciting area to gain insight in.This PhD project aims to develop a protocol to evaluate oil-water partition coefficients of active pharmaceutical compounds (APIs). This is an important investigation as the octanol-water partition model is still the foundation of obtaining an initial understanding of the solubility of APIs. The equilibrium concentration of a compound in each phase helps determine how lipophilic or hydrophilic a compound is (Bannan, et al., 2016). The water represents the aqueous media in the human body and the octanol represents biological membranes. Hence, the model is a surrogate for deciding the solubility of API in the body, and their transport across biological membranes. However, currently this is modelled through costly lab experiments. This project aims to create a suitable computational model to represent this experiment. A coarse-grained model will be used in which an API molecule is 'pulled' by an external force through the water-oil interface. During this process the energy barriers will be monitored to gain a deeper understanding of the transport process. Monitoring the energy barriers will also allow trends and relationships to be determined, if any. This should also help assess the current model used and determine why it may fail to fully represent the actual process of a drug distributing in the body. One of the main limitations of the experimental model is that it fails to account for non-equilibrium conditions. The computational model created should help gain insight into this part of the process.A large part of this project will involve the development of a suitable and accurate models. Therefore, the first stages of the project will involve doing an extensive literature review of the current models being used, such as atomistic models. A comparison will need to be made between these models, which will help inform the use of a coarse graining model in this project. There will also be other key milestones before the model can be fully set-up and simulated. These have been split into categories in the objectives, seen in the next section