Abstract

Quantitative structure-property/activity relationships (QSPRs/QSARs) are an important component of modern science. Validation of the QSPR/QSAR is the basis for applying. The system of self-consistent models is a new approach to validate QSPR/QSAR. The principle ‘QSAR is a random event’ means that an approach may be recognized as robust only if the statistical characteristics of models obtained by this approach for different splits (training/test) are reproduced. The above principle applies to the case of the nano-QSAR, also. Here, the cellular uptake of nanoparticles in pancreatic cancer cells examines as the endpoint. Groups of models for different splits (training/test) are compared. This comparison gives the possibility to formulate the system of self-consistent models as a way to assess the predictive potential for an arbitrary QSPR/QSAR and/or nano-QSPR/QSAR. The correlation intensity index (CII) has been tested as a tool to improve the quality of models for the cellular uptake of nanoparticles in pancreatic cancer cells (PaCa2). It has shown, that the CII can be useful, but only incorporating with the Index of ideality of correlation (IIC).

摘要

定量结构-性质/活性关系 (QSPRs/QSARs) 是现代科学的重要组成部分。QSPR/QSAR 的验证是申请的基础。自洽模型系统是一种验证QSPR/QSAR的新方法。“QSAR 是随机事件”的原则意味着只有在重现通过该方法针对不同拆分（训练/测试）获得的模型的统计特征时，才能认为该方法是稳健的。上述原理也适用于纳米 QSAR 的情况。在这里，胰腺癌细胞中纳米颗粒的细胞摄取被视为终点。比较不同分割（训练/测试）的模型组。这种比较提供了制定自洽模型系统的可能性，作为评估任意 QSPR/QSAR 和/或纳米 QSPR/QSAR 的预测潜力的一种方式。相关强度指数（CII ) 已被测试为提高胰腺癌细胞 (PaCa2) 中纳米颗粒细胞摄取模型质量的工具。它表明，CII可能有用，但只能与相关性理想指数 ( IIC ) 结合使用。