Improving the Accuracy of Implicit Solvents with a Physics-Guided Neural Network

利用物理引导神经网络提高隐式溶剂的准确性

基本信息

批准号：
10669809
负责人：
Negin Forouzesh
金额：
$ 18.25万
依托单位：
CALIFORNIA STATE UNIVERSITY LOS ANGELES
依托单位国家：
美国
项目类别：
财政年份：
2022
资助国家：
美国
起止时间：
2022-08-01 至 2026-05-31
项目状态：
未结题

来源：
https://reporter.nih.gov/project-details/10669809
关键词：
Acceleration Adherence Affinity Benchmarking Binding Binding Proteins Biological Biological Process Biological Sciences Cells Characteristics Chemicals Communicable Diseases Communities Complex Computer Assisted Computer Models Computer Simulation Computing Methodologies Coupling DNA biosynthesis Data Data Set Databases Drug Design Ensure Free Energy GTP-Binding Proteins Gene Expression Goals Health Human body Hybrids Interdisciplinary Study Ligands Machine Learning Measures Medicine Methods Modeling Modernization Modification Neural Network Simulation Outcomes Research Performance Physics Play Preparation Property Protein Conformation Proteins Protocols documentation Research Running Sampling Side Signal Transduction Signal Transduction Pathway Solvents Source Source Code Structure Study models System Techniques Testing Training Water Weight cost data cleaning data modeling data-driven model design drug candidate drug discovery drug-like compound experimental study flexibility foot graph neural network improved laboratory experiment loss of function molecular recognition neglect neural network novel novel therapeutics physical model prevent programs protein folding protonation prototype response simulation small molecule stem structural biology usability virtual screening web services

项目摘要

Project Summary/Abstract Drug discovery is one of the most challenging tasks in biological sciences; it takes about 10-15 years and $2 billion on average to discover a new drug. The main goal in drug discovery is identifying drug-like com- pounds (ligands) capable of modulating speciﬁc biological targets (proteins). One key feature of protein-ligand interactions is the binding free energy change, G, that occurs between the protein and the ligand upon the ligand's attachment. This physiochemical feature heavily dictates how strongly a protein and ligand interact and is particularly useful to understand for drug design. While wet-lab experiments accurately estimate G, they are signiﬁcantly slow, costly, and laborious. On the other hand, computational simulations enable signiﬁcantly faster estimation of G and shed light on the binding mechanism of various structures that could have been complicated to be examined otherwise. The implicit solvent framework, which treats solvent as a continuum with the dielectric and non-polar properties of water, offer much more efﬁcient estimation of G compared to other computational methodologies, such as alchemical free energy methods. Despite noticeable progress in implicit solvent modeling, serious concerns about its accuracy remain that stem from the underlying physi- cal approximations. This research will employ modern machine learning techniques to bridge the accuracy gap between a physics-based implicit solvent model and experimental references in terms of G calculations. In particular, experimental data will be integrated into a generalized Born (GB) implicit solvent model so that with adherence to the physical model, new structural features could improve the accuracy. In addition to the model accuracy, it is essential to retain interpretability (that accounts for the model simplicity) and transferability (that assures consistent performance on different datasets). To this end, a novel multi-objective loss function will be introduced that takes “accuracy”, “interpretability”, and “transferability” into consideration. Standard protein-ligand databases, benchmarks, and datasets will be used for designing the proposed hybrid model, including host-guest systems, SAMPL challenge benchmarks, PDBbind, and BindingDB. While some of these sources contain clean data, many require further post-processing to prepare for running the GB model. Careful data preparation will be performed by following standard protocols and via popular web services. The modular characteristics of the proposed physics-data model will allow for testing various ﬂavors of implicit solvent (physics-based model) and modiﬁcations to the proposed Graph Convolutional Network (data-driven model). This ﬂexibility of the hybrid model facilitates new interdisciplinary research between the classical physics-based and the modern data-driven ends. The ﬁnal source code and parameterized datasets will be available freely to the public. They could be incorporated into the high-throughput virtual screening of candidate drugs in the early stages of drug discovery. The outcome of this research will beneﬁt the biomolecular modeling community by providing an approach to build novel, accurate, and efﬁcient computational models for studying protein-ligand interactions.

项目摘要/摘要药物发现是生物科学中最挑战的任务之一。大约需要10 - 15年的时间平均20亿美元发现一种新药。药物发现的主要目的是确定类似药物的com- 磅（配体）能够调节特定的生物靶标（蛋白质）。蛋白质配体的一个关键特征相互作用是蛋白质和配体之间发生的结合自由能变化G 配体的依恋。这种生理化学特征在很大程度上决定了蛋白质和配体相互作用的强度和对于吸毒设计特别有用。当湿性实验准确估计G时，它们是非常慢，昂贵和费力。另一方面，计算模拟可以显着更快地估计G和Shed阐明可能是各种结构的结合机制复杂的检查，另外进行检查。隐式溶剂框架，将溶剂视为连续性与水的饮食和非极性特性相比，G相比提供了更多有效的G估计用于其他计算方法，例如酒精自由能法。尽管取得了明显的进步在隐式溶剂建模中，对其准确性的严重关注仍然是基本的物理学 CAL近似。这项研究将采用现代机器学习技术来弥合准确性差距基于物理学的隐式溶剂模型与G计算方面的实验参考之间。在特别的，实验数据将集成到广义诞生（GB）隐式溶剂模型中，以便与遵守物理模型，新的结构特征可以提高准确性。除了模型准确性，必须保留可解释性（解释模型简单性）和可转让性（确保在不同数据集上保持一致的性能）。为此，新型的多目标损失函数将是引入了考虑“准确性”，“可解释性”和“可转让性”的介绍。标准蛋白质配体数据库，基准和数据集将用于设计提出的混合模型，包括主机 - 网络系统，采样挑战基准，PDBBIND和BINDINGDB。这些来源中有一些包含清洁数据，许多需要进一步的后处理才能准备运行GB模型。仔细的数据准备将可以通过以下标准协议和流行的Web服务执行。模块化特征拟议的Physis-Data模型将允许测试隐式解决方案的各种大道（基于物理的模型）和对所提出的图形卷积网络（数据驱动模型）的修改。混合动力的这种功能模型促进了基于经典物理与现代数据驱动的新的跨学科研究结束。最终的源代码和参数化数据集将免费向公众提供。他们可能是在药物发现的早期阶段，纳入了候选药物的高通量虚拟筛查。这项研究的结果将通过提供一种建造方法来利用生物分子建模社区用于研究蛋白质 - 配体相互作用的新型，准确和有效的计算模型。