Not Drug-like, but Like Drugs: Cnidaria Natural Products.
Not Drug-like, but Like Drugs: Cnidaria Natural Products.
复制标题
DOI：
10.3390/md20010042
复制
发表时间：
2021-12-30
期刊：
影响因子：
5.4
通讯作者：
Johnson MP
中科院分区：
文献类型：
--
作者：
Laguionie-Marchais C;Allcock AL;Baker BJ;Conneely EA;Dietrick SG;Kearns F;McKeever K;Young RM;Sierra CA;Soldatou S;Woodcock HL;Johnson MP
刺胞动物门是天然产物的极好来源，鉴定出数千种代谢产物。其中许多尚未在生物测定中筛选。本研究的目的是探索5600刺胞藻天然产物的潜力（排除那些已知来自微生物共生体），使用基于化学空间，药物相似性，预测毒性和虚拟屏幕的系统方法。先前的药物类似措施：五法则、药物相似性的定量估计（QED）和相对药物相似性（RDL）基于相对少量的分子性质。我们增强了这种方法使用参考药物和毒素数据集定义的51个预测的分子特性。刺胞藻天然产物在这个化学空间中与药物和毒素重叠，尽管多变量测试表明各组之间存在一些差异。在已建立的药物相似性措施方面，刺胞藻天然产品的QED和RDL评分通常低于药物，超过至少一个五分法则阈值的代谢产物的患病率较高。然而，包括预测毒性（ADMET评分）的药物相似性指数发现，刺胞藻天然产品比药物更有利。单个刺胞藻天然产物到多元化学空间中药物分布中心的距离的测量与RDL、ADMET分数和五例外规则的数量有关。这种多变量相似性度量与相同代谢物的QED评分呈负相关，表明不同的方法捕获了单个代谢物药物相似性的不同方面。不同药物相似性度量的对比有助于总结刺胞藻天然产物数据集的药物潜力范围。最有利的代谢产物在210-265 Da左右，通常是倍半萜，具有中等程度的复杂性。针对癌症相关靶标的虚拟筛选发现了广泛的亲和力证据，其中19%的刺胞动物天然产物的Glide评分<-7。
Phylum Cnidaria has been an excellent source of natural products, with thousands of metabolites identified. Many of these have not been screened in bioassays. The aim of this study was to explore the potential of 5600 Cnidaria natural products (after excluding those known to derive from microbial symbionts), using a systematic approach based on chemical space, drug-likeness, predicted toxicity, and virtual screens. Previous drug-likeness measures: the rule-of-five, quantitative estimate of drug-likeness (QED), and relative drug likelihoods (RDL) are based on a relatively small number of molecular properties. We augmented this approach using reference drug and toxin data sets defined for 51 predicted molecular properties. Cnidaria natural products overlap with drugs and toxins in this chemical space, although a multivariate test suggests that there are some differences between the groups. In terms of the established drug-likeness measures, Cnidaria natural products have generally lower QED and RDL scores than drugs, with a higher prevalence of metabolites that exceed at least one rule-of-five threshold. An index of drug-likeness that includes predicted toxicity (ADMET-score), however, found that Cnidaria natural products were more favourable than drugs. A measure of the distance of individual Cnidaria natural products to the centre of the drug distribution in multivariate chemical space was related to RDL, ADMET-score, and the number of rule-of-five exceptions. This multivariate similarity measure was negatively correlated with the QED score for the same metabolite, suggesting that the different approaches capture different aspects of the drug-likeness of individual metabolites. The contrasting of different drug similarity measures can help summarise the range of drug potential in the Cnidaria natural product data set. The most favourable metabolites were around 210–265 Da, quite often sesquiterpenes, with a moderate degree of complexity. Virtual screening against cancer-relevant targets found wide evidence of affinities, with Glide scores <−7 in 19% of the Cnidaria natural products.
登录
查看更多内容
影响因子：
21.8
作者：
通讯作者：
--
影响因子：
14.9
作者：
Lim E;Pon A;Djoumbou Y;Knox C;Shrivastava S;Guo AC;Neveu V;Wishart DS
通讯作者：
Wishart DS
影响因子：
1.8
作者：
Lin, Yun-Sheng;Fazary, Ahmed Eid;Shen, Ya-Ching
通讯作者：
Shen, Ya-Ching
DOI：
10.1021/ci0200467
发表时间：
2003-01-01
期刊：
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES
影响因子：
--
作者：
Feher, M;Schmidt, JM
通讯作者：
Schmidt, JM
影响因子：
4.6
作者：
Daina A;Michielin O;Zoete V
通讯作者：
Zoete V