The objective of this study was to correlate the binding of drugs on a very popular nanoparticulate polymeric matrix; PLGA nanoparticles with their main constitutional, electronic and physico-chemical descriptors. Gaussian Processes (GPs) was the artificial intelligence machine learning method that was utilized to fulfil this task. The method could successfully model the results where optimum values of the investigated descriptors of the loaded drugs were deduced. A percentage bias of 12.68% ± 2.1 was obtained in predicting the binding energies of a group of test drugs. As a conclusion, GPs could successfully model the drugs-PLGA interactions associated with a good predicting power. The GPs-predicted binding energies (ΔG) can easily be projected to the drugs loading as was previously proven. Adopting the “Pharmaceutics Informatics” approach can save the pharmaceutical industry and the drug delivery scientists a lot of exerted resources, efforts and time.

这项研究的目的是将药物与非常流行的纳米颗粒聚合物基质的结合联系起来；PLGA 纳米粒子及其主要的构成、电子和物理化学描述符。高斯过程 (GP) 是用于完成此任务的人工智能机器学习方法。该方法可以成功地对结果进行建模，推导出所研究的装载药物描述符的最佳值。在预测一组测试药物的结合能时获得了 12.68% ± 2.1 的百分比偏差。总之，GP 可以成功地模拟与良好预测能力相关的药物-PLGA 相互作用。如前所述，GP 预测的结合能 (ΔG) 可以很容易地投射到药物负载上。