摘要
目的探讨白细胞中循环miRNA在精神分裂症SZ的表达水平及其诊断价值。方法收集90例SZ患者作为疾病组,90例双相情感障碍(BPAD)患者作为病例对照组,90例体检健康者作为健康对照组,采用实时荧光定量PCR检测各组白细胞中8个miRNA(miR-31-5p、miR-99b-5p、miR-107、miR-134-5p、miR-487b-3p、miR-181b-3p、miR-431-5p和miR-433-5p)的表达水平,并进行统计学分析,以验证差异表达的miRNA;绘制受试者工作曲线(ROC)曲线,建立人工神经网络(ANN)诊断模型,并分析其诊断价值。结果 SZ患者存在6个显著低表达的miRNA,其ROC曲线下面积(AUCROC)分别为miR-31-5p:0.931,miR-487b-3p:0.887,miR-99b-5p:0.869,miR-431-3p:0.763,miR-134-3p:0.756,miR-107:0.721;在ANN模型中,最优模型的AUCROC为0.960,敏感性86.7%,特异性95.6%。结论白细胞中6个差异表达的miRNA可能作为SZ诊断的非侵袭性生物学标志物,而ANN诊断模型可以提高miRNA对SZ的诊断价值。
Objective To investigate the expression level of circulating miRNA in WBC of the patients with schizophrenia and its diagnostic value. Methods The expression levels of eight miRNAs,including miR-31-5p,miR-99b-5p,miR-107,miR-134-5p,miR-487b-3p,miR-181b-3p,miR-431-5p and miR-433-5p,in WBCs of peripheral blood samples from 90 patients with schizophrenia,90 patients with bipolar disorder,and 90 normal controls were detected by real-time quantitative PCR. Then,the differentially expressed miRNAs were validated based on statistical analysis. The diagnostic value of these miRNAs was evaluated by receiver operating curve（ ROC） and the established artificial neural network（ ANN） model. Results The expression levels of 6 miRNAs were significantly lower in schizophrenia than those in bipolar disorder and normal controls,and their AUC^ROCwere 0. 931 for miR-31-5p,0. 887 for miR-487b-3p,0. 869 for miR-99b-5p,0. 763 for miR-431-3p,0. 756 for miR-134-3p,and 0. 721 for miR-107,respectively. The AUC^ROC,sensitivity and specificity of the optimal ANN model were 0. 960,86. 7% and 95. 6%,respectively. Conclusion Six differentially expressed miRNAs in WBC may be used as noninvasive biomarkers for the diagnosis of schizophrenia,and ANN model can improve the diagnostic value of miRNA in schizophrenia.
出处
《临床检验杂志》
CAS
CSCD
2016年第4期247-250,共4页
Chinese Journal of Clinical Laboratory Science
基金
国家863计划(2006AA02090407)
