Matrix metalloproteinase-2 (MMP-2) is a lucrative therapeutic target as far as anticancer drug discovery is concerned. Overexpression of MMP-2 is found to facilitate tumour propagation through the involvement of vascular endothelial growth factor (VEGF). However, even after different techniques, finding a target-specific MMP-2 inhibitor with respectable pharmacodynamic properties is still a challenging task. Regression-dependent quantitative structure–activity relationship (QSAR) strategies might be among the possible drug design methods to explore the essential structural features that would be valuable to find a suitable MMP-2 inhibitor. In this paper, 72 molecules were explored using the PaDEL descriptors and stepwise multiple linear regression (S-MLR). The partial least squares (PLS) method was also used to create a viable statistical model with an acceptable metric related to these models. The final statistical models were formed with statistical parameters within acceptable range (r² = 0.797, Q² = 0.725 and r²_pred = 0.643 for the MLR model, and r² = 0.780, Q² = 0.685 and r²_pred = 0.666 for the PLS model). The models were analysed and compared with those already published on the same endpoint.

就抗癌药物的发现而言，基质金属蛋白酶2（MMP-2）是一种有利可图的治疗靶标。发现MMP-2的过表达通过血管内皮生长因子（VEGF）的参与促进肿瘤的扩散。然而，即使采用不同的技术，寻找具有可观的药效特性的靶标特异性MMP-2抑制剂仍然是一项艰巨的任务。依赖于回归的定量结构-活性关系（QSAR）策略可能是探索可能的基本结构特征的药物设计方法之一，这些结构特征对于找到合适的MMP-2抑制剂将是有价值的。在本文中，使用PaDEL描述子和逐步多元线性回归（S-MLR）探索了72个分子。偏最小二乘（PLS）方法也用于创建可行的统计模型，并具有与这些模型相关的可接受度量。最终统计模型的统计参数在可接受范围内（对于MLR模型，r ² = 0.797，Q ² = 0.725，r ² _pred = 0.643，而对于PLS模型，r ² = 0.780，Q ² = 0.685，r ² _pred = 0.666。对模型进行了分析，并与已经在同一端点上发布的模型进行了比较。