Discovery Sciences Computational Sciences, AstraZeneca R&D, Pepparedsleden 1, S-43183 Mölndal, Sweden.
J Chem Inf Model. 2012 Jul 23;52(7):1777-86. doi: 10.1021/ci300157m. Epub 2012 Jul 2.
An early stage drug discovery project needs to identify a number of chemically diverse and attractive compounds. These hit compounds are typically found through high-throughput screening campaigns. The diversity of the chemical libraries used in screening is therefore important. In this study, we describe a virtual high-throughput screening system called Virtual Library. The system automatically "recycles" validated synthetic protocols and available starting materials to generate a large number of virtual compound libraries, and allows for fast searches in the generated libraries using a 2D fingerprint based screening method. Virtual Library links the returned virtual hit compounds back to experimental protocols to quickly assess the synthetic accessibility of the hits. The system can be used as an idea generator for library design to enrich the screening collection and to explore the structure-activity landscape around a specific active compound.
一个早期的药物发现项目需要确定许多化学多样性和有吸引力的化合物。这些命中化合物通常是通过高通量筛选活动发现的。因此，筛选中使用的化学文库的多样性很重要。在这项研究中，我们描述了一个名为 Virtual Library 的虚拟高通量筛选系统。该系统自动“回收”经过验证的合成方案和可用的起始材料，以生成大量虚拟化合物库，并允许使用基于 2D 指纹的筛选方法在生成的库中进行快速搜索。Virtual Library 将返回的虚拟命中化合物与实验方案联系起来，以便快速评估命中化合物的合成可及性。该系统可用作库设计的创意生成器，以丰富筛选集并探索特定活性化合物周围的结构活性景观。
