摘要目的 利用胰岛转录组数据构建共表达网络,鉴定与2型糖尿病发病相关的重基因及其作用机制.方法 从NCBI的GEO数据库中下载芯片数据GSE38642,包括2型糖尿病样本9例、糖尿病前期样本9例(6%≤HbA1C<6.5%)、正常对照样本31例(HbA1C<6%).用R语言权重基因共表达网络分析( weighted gene co-expression network analysis, WGCNA)包构建基因共表达网络并划分模块,选择与临床指标相关的基因模块,用R语言的GeneAnswers包进行GO分析和KEGG通路分析,并结合TRANSFAC数据库进行上游转录因子分析,选择模块的枢纽基因和上游转录因子作为候选重基因.结果 基因共表达网络包括34个模块. Green模块与HbA1C%呈正相关(R＝0.47, P＝1×10－4),GO富集分析结果主为细胞粘附、细胞外基质降解等,KEGG通路为胰腺分泌、粘着斑等. Green模块的枢纽基因有3个,分别为ITGA6、ZAK和YBX3. Brown模块与HbA1C呈负相关(R＝0.46, P＝1×10－4),GO富集分析结果为突触、跨膜转运蛋白活性和运输等,KEGG通路为胰岛素分泌、多巴胺能突触等,上游转录因子PAX6、REST和PDX1可能在疾病中发挥了重作用. Brown模块的枢纽基因包括SLC4A10、ELAVL4、SYT14等30个基因.既往研究示这些基因可能在2型糖尿病的发病机制中起了重作用.结论 通过构建基因共表达网络的方法能够从转录组数据中挖掘到2型糖尿病发病相关的重基因,为疾病研究供新的候选基因和分子机制.

abstractsObjective To identify potential molecule targets of type 2 diabetes using weighted gene co-expression network analysis. Methods Microarray data of type 2 diabetes (GSE38642) were downloaded from Gene Expression Omnibus of NCBI, including 9 type 2 diabetic patients, 9 pre-diabetic patients ( 6%≤HbA1C<6.5%), and 31 normal controls (HbA1C<6%). Using weighted gene co-expression network analysis (WGCNA) package in R, the weighted gene co-expression network was built and significant modules related to clinical traits were identified. Then, functional and pathway enrichment analysis were conducted for genes in the most significant modules using GeneAnswers package in R. Upstream transcription factor enrichment analysis were conducted using TRANSFAC database. The hub genes and upstream transcription factors were selected as potential molecule targets of type 2 diabetes. Results 34 modules were identified in the co-expression network. Green module was positive correlated with HbA1C(R＝0.47, P＝1×10－4). The enriched functions were cell adhesion, extracellular matrix disassembly, etc. The enriched KEGG pathways were Pancreatic secretion, Focal adhesion, etc. ITGA6, ZAK, and YBX3 are hub genes of Green module. Brown module was negative correlated with HbA1C(R＝0.46, P＝1×10－4). The enriched functions were synapse, transmembrane transporter activity, etc. The enriched KEGG pathways were Insulin secretion, Dopaminergic synapse, etc. The upstream transcription factors PAX6, REST, and PDX1 of Brown module might play important roles. 30 hub genes, including SLC4A10, ELAVL4, and SYT14, were identified in Brown module. The relationships between these genes and type 2 diabetes were confirmed by previously published studies. Conclusion Important genes related to type 2 diabetes can be filtered out from transcriptome profiles using gene co-expression analysis. Our finding might provide a novel insight into the underlying molecular mechanism of type 2 diabetes.