Abstract：
Targeting cancer cells through drug-based treatment or combination therapy protocols involving chemical compounds can be challenging due to multiple factors, including their resistance to bioactive compounds and the potential of drugs to damage healthy cells. This study aims to investigate the relationship between the structure of novel sulfur-containing shikonin oxime compounds and the corresponding cytotoxicity against four cancer types, namely colon, gastric, liver, and breast cancers, through computational chemistry tools. This investigation is suggested to help build insights into how the structure of the compounds influences their activity and understand the mechanisms behind it and subsequently might be used in multi-cancer drug design process to propose novel optimized compounds that potentially exhibit the desired activity. The findings showed that the cytotoxic activity against the four cancer types was accurately predictable (R2 > 0.7, NRMSE <20%) by a combination of search and machine learning algorithms, based on the information on the structure of the compounds, including their lipophilicity, surface area, and volume. Overall, this study is supposed to play a crucial role in effective multi-cancer drug design in cancer research areas.
摘要：
由于多种因素，通过基于药物的治疗或涉及化合物的联合治疗方案靶向癌细胞可能具有挑战性。包括它们对生物活性化合物的抗性和药物损害健康细胞的潜力。本研究旨在探讨新型含硫紫草素肟化合物的结构与对四种癌症类型的细胞毒性之间的关系。即结肠，胃,肝脏,和乳腺癌，通过计算化学工具。这项研究被认为有助于深入了解化合物的结构如何影响其活性，并了解其背后的机制，随后可用于多种癌症药物设计过程，以提出可能表现出所需活性的新型优化化合物。研究结果表明，通过搜索和机器学习算法的组合，对四种癌症类型的细胞毒性活性是准确预测的（R2>0.7，NRMSE<20％），根据有关化合物结构的信息，包括它们的亲脂性，表面积,和音量。总的来说,这项研究应该在癌症研究领域有效的多种癌症药物设计中发挥关键作用。
