肿瘤干细胞特征及干性指数机器算法的研究进展

Progress on stemness characteristics of cancer stem cells and machine algorithms for stemness index

目的就肿瘤干细胞特征及肿瘤干性指数相关算法进行综述,以期更好地指导肿瘤患者的精准化治疗,从而改善肿瘤患者预后。方法以“cancer stem cell、stemness index、stemness characteristics、machine algorithm、OCLR、ssGSEA、REOs”等为关键词检索PubMed和Web of Science数据库2011-05-01-2022-01-01相关文献。纳入标准:(1)肿瘤干细胞特征相关研究;(2)肿瘤干细胞与肿瘤的发生、侵袭、转移和耐药相关研究;(3)干性指数算法;(4)干性指数算法的应用。排除标准:(1)数据不完整的文献;(2)实验结果重复文献;(3)数据和结果陈旧文献。最终纳入分析文献41篇。结果肿瘤干细胞具有自我更新能力,与肿瘤细胞的侵袭、转移、耐药和复发相关。目前肿瘤干性指数的机器算法主要有一类逻辑回归(OCLR)算法、单样本基因集富集分析(ssGSEA)算法和相对表达秩序关系(REOs)算法。结论肿瘤干细胞的功能特征定性描述了肿瘤干性,而通过机器算法得到的干性指数,则可定量描述肿瘤干性,两者相辅相成,有助于对肿瘤干细胞的作用机制进行深入研究。

Objective To review the characteristics of tumor stem cells and the algorithm related to the tumor stemness index,in order to better guide the precise treatment of tumor patients,improve the prognosis of tumor patients.Methods Based on PubMed and Web of Science databases,literatures published from May 1,2011 to January 1,2022 were retrieved with the keywords such as"cancer stem cells,stemness index,stemness characteristics,machine algorithm,OCLR,ssGSEA,REOs".Inclusion criteria:(1)Researches on characteristics of cancer stem cells.(2)Researches on cancer stem cells and tumor occurrence,invasion,metastasis and drug resistance.(3)Stemness index algorithm.(4)Application of stemness index algorithm.Exclusion criteria:(1)Literatures with incomplete data.(2)Literatures with duplicate experimental results.(3)Literature with outdated data and results.Finally,41 articles were included in the analysis.Results Cancer stem cells have self-renewal ability and are related to tumor cell invasion,metastasis,drug resistance and recurrence.At present,the machine algorithms of tumor stemness index mainly include one-class logistic regression(OCLR)algorithm,single-sample gene set enrichment analysis(ssGSEA)algorithm and relative expression orderings(REOs)algorithm.Conclusion The functional characteristics of tumor stem cells qualitatively describe the tumor stemness,while the stemness index obtained by machine algorithm can quantitatively describe the tumor stemness.