CAS in Crystallography and Biophysics, University of Madras, Guindy Campus, Chennai - 600025, India.
Department of Biotechnology, SRM Institute of Science and Technology, Kattankulathur - 603203, Kanchipuram District, Tamilnadu, India.
Curr Top Med Chem. 2020;20(19):1761-1770. doi: 10.2174/1568026620666200622150003.
Structure-based drug design is a wide area of identification of selective inhibitors of a target of interest. From the time of the availability of three dimensional structure of the drug targets, mostly the proteins, many computational methods had emerged to address the challenges associated with drug design process. Particularly, drug-likeness, druggability of the target protein, specificity, off-target binding, etc., are the important factors to determine the efficacy of new chemical inhibitors.
The aim of the present research was to improve the drug design strategies in field of design of novel inhibitors with respect to specific target protein in disease pathology. Recent statistical machine learning methods applied for structural and chemical data analysis had been elaborated in current drug design field.
As the size of the biological data shows a continuous growth, new computational algorithms and analytical methods are being developed with different objectives. It covers a wide area, from protein structure prediction to drug toxicity prediction. Moreover, the computational methods are available to analyze the structural data of varying types and sizes of which, most of the semi-empirical force field and quantum mechanics based molecular modeling methods showed a proven accuracy towards analysing small structural data sets while statistics based methods such as machine learning, QSAR and other specific data analytics methods are robust for large scale data analysis.
In this present study, the background has been reviewed for new drug lead development with respect specific drug targets of interest. Overall approach of both the extreme methods were also used to demonstrate with the plausible outcome.
In this chapter, we focus on the recent developments in the structure-based drug design using advanced molecular modeling techniques in conjunction with machine learning and other data analytics methods. Natural products based drug discovery is also discussed.
基于结构的药物设计是识别目标感兴趣的选择性抑制剂的一个广泛领域。自从药物靶点（主要是蛋白质）的三维结构可用以来，已经出现了许多计算方法来解决与药物设计过程相关的挑战。特别是，药物相似性、目标蛋白质的可成药性、特异性、脱靶结合等，是决定新化学抑制剂疗效的重要因素。
本研究旨在针对疾病病理学中特定的靶标蛋白，改进新药设计策略，设计新型抑制剂。目前药物设计领域已经详细阐述了最近应用于结构和化学数据分析的统计机器学习方法。
随着生物数据量的不断增长，不同目标的新计算算法和分析方法正在被开发。它涵盖了从蛋白质结构预测到药物毒性预测的广泛领域。此外，还有计算方法可用于分析不同类型和大小的结构数据，其中大多数半经验力场和基于量子力学的分子建模方法在分析小型结构数据集方面表现出了良好的准确性，而基于统计学的方法，如机器学习、QSAR 和其他特定数据分析方法，则适用于大规模数据分析。
在本研究中，针对特定的药物靶标，对新药先导物的开发进行了背景回顾。还使用了两种极端方法的综合方法来展示合理的结果。
在这一章中，我们专注于使用先进的分子建模技术结合机器学习和其他数据分析方法进行基于结构的药物设计的最新进展。还讨论了基于天然产物的药物发现。
