基于核苷酸序列的探针水平数据分析

微阵列实验是一个复杂的多步骤的实验过程,不确定性存在于实验的每一个步骤中,导致最后得到的实验结果中包含了一些数据噪声。为了从这些含有噪声的数据中得到更多有意义的生物信息,很多算法相继被提出来计算基因表达值。目前流行的mmgMOS模型提高了芯片数据分析的准确性,但是该模型的主要缺点是其参数值φ在整个数据集上是唯一不变的,单一的值不能代表不同探针的真实信号。本文对mmgMOS模型中的参数值φ进行改进,从而进一步提高后续寻找差异基因的准确率。

Analysis of Probe-Level Gene Expression Based on Nucleotide Sequence

Microarray experiment is a complicated multi-step procedure,and since variability can be introduced at each experimental stage that caused some noisy gene expression data in the experiment.In order to obtain much valuable information from the data,many algorithms have been proposed for estimating gene expression level.At present,the most popular model,mmgMOS can improve the accuracy of microarray data analysis,but has a deficiency that the value of the parameter φ used throughout the whole data set is fixed,while the fixed single value cannot represent real signals from various probes.An alternative way for appropriate parameter setting is proposed to improve the accuracy in finding differential gene expression level.