Scaffold Distributions in Bioactive Molecules, Clinical Trials Compounds, and Drugs
Scaffold Distributions in Bioactive Molecules, Clinical Trials Compounds, and Drugs
复制标题
DOI：
10.1002/cmdc.200900419
复制
发表时间：
2010-02-01
期刊：
影响因子：
3.4
通讯作者：
Bajorath, Juergen
中科院分区：
文献类型：
--
作者：
Hu, Ye;Bajorath, Juergen
包括支架(核心结构)[1-4]和不同来源和大小的片段[5-8]在内的分子构建块在寻找靶类导向的结构基序[9-11]和基于片段的药物发现中得到了深入的研究。[12-15]在这些研究的背景下，常常有可能将支架、片段或片段的组合与特定或多种生物学活性联系起来。[4-11]这些研究中的大多数都是在频率分析的基础上调查片段在生物相关化合物中的分布。对已知药物[1]或针对不同目标类别的化合物[10，11]中存在的支架进行了分析，以评估区分药物与非药物或某些药物类别的特征的结构特征。这些面向结构的研究在概念上与其他关于类铅或类药物化合物性质的研究有关，这些研究主要集中在借助各种分子描述符分析分子性质分布。[16-19]综合起来，这些和其他研究在很大程度上有助于阐明各种生物活性的结构特征，并识别与药物或铅相似的分子性质分布。我们一直感兴趣的是从不同的角度分析支架分布，即为了更好地了解药物开发不同阶段化合物的结构特征如何比较。因此，我们对代表药物发现不同阶段的三组化合物进行了比较分子支架分析：生物活性分子(HITS或先导)、临床试验中的化合物以及注册/批准的药物。通过这一分析，我们试图探索几个问题。例如，在不同的发展阶段，支架人群的构成是否会有显著差异？某些支架是否会优先出现在早期，而不是晚期化合物或药物中？或者某些支架会一直存在于这些类型的化合物中吗？显然，在某种程度上，这样的问题是由候选药物的高临床损失率引起的。[20，21]最初，我们组装了合适的化合物数据集。作为一个命中和线索池，我们从BindingDB中检索到了针对人类目标的所有活性分子，[22]它包含
Molecular building blocks including scaffolds (core structures)[1–4] and fragments of varying origin and size [5–8] have been intensely investigated in the search for target-class-directed structural motifs [9–11] and in fragment-based drug discovery.[12–15] In the context of these studies, it has often been possible to associate scaffolds, fragments, or combinations of fragments with specific or multiple biological activities.[4–11] The majority of these studies have surveyed distributions of fragments in biologically relevant compounds on the basis of frequency analysis.[5–10] Furthermore, scaffolds present in known drugs [1] or compounds directed against different target classes [10, 11] have been analyzed in order to evaluate structural features that distinguish drugs from non-drugs or that are characteristic of certain drug classes. These structure-oriented investigations are conceptually related to other studies of leadlike or drug-like compound character that have predominantly focused on analyzing molecular property distributions with the aid of various molecular descriptors.[16–19] Taken together, these and other studies have substantially aided in elucidating structural signatures of various biological activities and in identifying molecular property distributions consistent with drug-or lead-likeness.We have been interested in analyzing scaffold distributions from a different perspective, that is, in order to better understand how structural features in compounds at different stages of pharmaceutical development might compare. Therefore, we carried out a comparative molecular scaffold analysis of three sets of compounds representing different stages in drug discovery: biologically active molecules (hits or leads), compounds in clinical trials, and registered/approved drugs. With this analysis, we attempted to explore several questions. For example, would there be notable differences in the composition of scaffold populations at different development stages? Might some scaffolds preferentially occur in early-but not latestage compounds or drugs? Or would certain scaffolds be consistently found in these types of compounds? Clearly, such questions are, to some extent, inspired by the high clinical attrition rates of drug candidates.[20, 21] Initially, we assembled suitable compound data sets. As a pool of hits and leads, we retrieved all active molecules directed against human targets from BindingDB,[22] which contains