应用数据挖掘及深度学习技术探索皮肤鳞状细胞癌的治疗靶点及药物  

Identification of drug compounds for cutaneous squamous cell carcinoma:drug discovery based on text mining and deep learning

作　　者：潘昱妍 陈志炜 刘家祺 Pan Yuyan;Chen Zhiwei;Liu Jiaqi(Department of Plastic Surgery,Zhongshan Hospital,Fudan University,Shanghai 200032,China;Big Data and Artificial Intelligence Center,Zhongshan Hospital,Fudan University,Shanghai 200032,China;Artificial Intelligence Center for Plastic Surgery and Cutaneous Soft Tissue Cancers,Zhongshan Hospital,Fudan University,Shanghai 200032,China)

机构地区：[1]复旦大学附属中山医院整形外科,上海200032 [2]复旦大学附属中山医院大数据与人工智能中心,上海200032 [3]复旦大学附属中山医院整形与皮肤软组织肿瘤人工智能中心,上海200032

出　　处：《中华整形外科杂志》2022年第11期1210-1221,共12页Chinese Journal of Plastic Surgery

基　　金：国家自然科学基金青年科学基金(81802724, 82102333)。

摘　　要：目的使用计算机工具和公开数据库挖掘与皮肤鳞状细胞癌(cSCC)相关的基因和信号通路, 并通过深度学习模型探索治疗cSCC的靶点及药物。方法通过文本挖掘和GeneCodis找出与cSCC高度相关的基因;使用STRING和Cytoscape进行蛋白质-蛋白质相互作用分析;通过DGIdb数据库基于药物-基因相互作用分析, 得到候选药物;利用药物-靶点相互作用深度学习模型DeepPurpose, 采用深度学习算法, 在药靶相关性的基础上进一步对药物-靶点亲和力进行预测, 并给出与目标靶点亲和力较高的部分药物推荐。结果通过文本挖掘识别出与cSCC相关的121个基因;基因富集分析中产生了与10个信号通路有关的11个基因和54个靶向药物。其中, 主要通路包括"pathways in cancer"(癌症相关信号通路)、"MAPK signaling pathway"(MAPK信号通路)、"ErbB signaling pathway"(ErbB信号通路);主要基因包括TP53、MDM2、CCND1、CDKN2A、HRAS、EGFR、MYC、ERBB2、AKT1、STAT3和SRC。通过DeepPurpose得到34个最终药物, 包括11个化疗药物、17个酪氨酸激酶抑制剂、4个PI3K/AKT/mTOR抑制剂、1个丝裂原活化蛋白激酶抑制剂和维生素A酸。结论使用计算机工具和深度学习模型有望成为一种新的探索靶向cSCC基因药物的有效方法。Objective To determine the genes and molecular pathways associated with cutaneous squamous cell carcinoma(cSCC)by using computational tools and machine learning models,and to explore drugs targeting the relevant genes for cSCC treatment.Methods Text mining and GeneCodis were used to mine genes which were highly related to cSCC.Protein-protein interaction analysis was performed by using STRING and Cytoscape.Gene-drug interaction analysis was performed by using the DGIdb database.Drug-target interaction prediction was performed by using deep learning model,DeepPurpose,through which candidate drugs with the highest predicted binding affinity were finally obtained.Results Our analysis identified 121 genes related to cSCC from the text mining searches.Gene enrichment analysis yielded 11 genes representing 10 pathways,targetable by a total of 54 drugs as possible drug treatments for cSCC.Among them,the main pathways included"pathways in cancer","MAPK signaling pathway"and"ErbB signaling pathway",while the genes included TP53,MDM2,CCND1,CDKN2A,HRAS,EGFR,MYC,ERBB2,AKT1,STAT3 and SRC.DeepPurpose recommended 34 drugs as the final drug list,including 11 chemotherapy agents,17 tyrosine kinase inhibitors(TKIs),4 PI3K/AKT/mTOR inhibitors,1 MAPK inhibitors and acitretin.Conclusions Drug discovery using in silico data mining and deep learning algorithm may be a potential powerful and effective way to identify drugs targeting the genes related to cSCC.

关 键 词：皮肤肿瘤 数据挖掘 深度学习 药物疗法 药物基因相互作用分析 

分 类 号：R739.5[医药卫生—肿瘤]