A set of 23 steroidal 1,2,4,5-tetraoxane analogues were studied using quantum-chemical method (B3LYP/6-31 G*) and multivariate analyses (PCA, HCA, KNN and SIMCA) in order to calculate the properties and correlate them with antimalarial activity (log RA) against Plasmodium falciparum clone D-6 from Sierra Leone. PCA results indicated 99.94% of the total variance and it was possible to divide the compounds into two classes: less and more active. Descriptors responsible for separating were: highest occupied molecular orbital energy (HOMO), bond length (O1-O2), Mulliken electronegativity (χ) and Bond information content (BIC0). We use HCA, KNN and SIMCA to explain relationships between molecular properties and biological activity of a training set and to predict antimalarial activity (log RA) of 13 compounds (#24-36) with unknown biological activity. We apply molecular docking simulations to identify intermolecular interactions with a selected biological target. The results obtained in multivariate analysis aided in the understanding of the activity of the new compound's design (#24-36). Thus, through chemometric analyses and docking molecular study, we propose theoretical synthetic routes for the most promising compounds 28, 30, 32 and 36 that can proceed to synthesis steps and in vitro and in vivo assays.

使用量子化学方法（B3LYP / 6-31 G *）和多元分析（PCA，HCA，KNN和SIMCA）研究了23种甾体1,2,4,5-四恶烷类似物，以计算其性质和使它们与抗恶性疟原虫的抗疟活性（log RA）相关从塞拉利昂克隆D-6。PCA结果表明总差异为99.94％，可以将化合物分为两类：活性较低和活性较高。负责分离的描述词是：最高占据分子轨道能（HOMO），键长（O1-O2），穆里肯电负性（χ）和键信息含量（BIC0）。我们使用HCA，KNN和SIMCA来解释训练集的分子特性和生物学活性之间的关系，并预测抗疟活性（log RA）的13种具有未知生物活性的化合物（＃24-36）。我们应用分子对接模拟来识别与选定的生物学目标的分子间相互作用。在多变量分析中获得的结果有助于理解新化合物设计的活性（＃24-36）。因此，通过化学计量分析和对接分子研究，我们提出了最有前途的化合物28、30、32和36的理论合成路线，这些路线可以进行合成步骤以及体外和体内测定。