Conditional estimation of local pooled dispersion parameter in small-sample RNA-Seq data improves differential expression test
Conditional estimation of local pooled dispersion parameter in small-sample RNA-Seq data improves differential expression test
复制标题
小样本 RNA-Seq 数据中局部合并分散参数的条件估计改进了差异表达测试
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
T. Park
中科院分区：
文献类型：
--
作者：
Jungsoo Gim;Sungho Won;T. Park
转录组学研究中的高通量测序技术有助于理解基因调控机制及其细胞功能，但也增加了对准确统计方法来评估实验之间定量差异的需求。人们已经开发了许多方法来解释计数数据的具体情况：非正态性、方差对平均值的依赖性以及小样本量。其中，典型实验中样本数量较少仍然是一个挑战。在这里，我们提出了一种使用局部合并离散参数的条件估计来对计数数据进行差异分析的方法。使用模拟和真实数据集对我们提出的方法在差异基因表达分析方面进行综合评估表明，在控制错误发现率的同时，该方法比其他现有方法更强大。通过引入局部合并分散参数的条件估计，我们成功克服了小功率的限制，并实现了针对少量样本的差异表达测试的强大定量分析。
High throughput sequencing technology in transcriptomics studies contribute to the understanding of gene regulation mechanism and its cellular function, but also increases a need for accurate statistical methods to assess quantitative differences between experiments. Many methods have been developed to account for the specifics of count data: non-normality, a dependence of the variance on the mean, and small sample size. Among them, the small number of samples in typical experiments is still a challenge. Here we present a method for differential analysis of count data, using conditional estimation of local pooled dispersion parameters. A comprehensive evaluation of our proposed method in the aspect of differential gene expression analysis using both simulated and real data sets shows that the proposed method is more powerful than other existing methods while controlling the false discovery rates. By introducing conditional estimation of local pooled dispersion parameters, we successfully overcome the limitation of small power and enable a powerful quantitative analysis focused on differential expression test with the small number of samples.
影响因子：
7
作者：
Blekhman, Ran;Marioni, John C.;Gilad, Yoav
通讯作者：
Gilad, Yoav