【摘要】 目的基于微阵列表达数据,探索筛选差异表达基因的有效方法。方法在传统t检验和方差分析的基础上,通过对重复实验数据的随机组合,构建秩统计量和期望秩统计量,通过调节二者的差值作为阈值,计算对应假阳性率(FPR),筛选差异表达基因。结果将本方法应用于模拟数据集和真实微阵列表达数据上,均取得较好的效果。结论本方法适用于微阵列表达数据,进行差异表达基因的筛选,并且比传统的方法更具灵活性。更多还原

【Abstract】 Objective To search an effective method on screening significant genes based on microarray data.Methods Based on conventional t test,ANOVA and random combined microarray data,the rank statistic and expected rank statistic were constructed.Though adjusting the difference of the two statistics and taking it as a threshold,the corresponding FPR(false positive rate) was computed and the significant genes were screened.Results With the method used on simulative data and real microarray data,idealized operation result was made.Conclusion The method is suitable to analyze microarray data and screen significant genes.Besides,the method is more flexible than conventional methods.更多还原