Whether traditional multi-cellular transcriptome sequencing (bulk RNA-seq) or single cell transcriptome sequencing (scRNA-seq), differential expression analysis (differential expression analysis) is the basic method of comparing two different samples of the similarities and differences of gene expression is obtained a set of samples with respect to another set of samples significantly upregulated expression (up-regulated) genes and down-regulation (down-regulated), which can be further study of these differentially expressed genes, including enriched passage (pathway) or biological procedure (biological process).
Due to the limitations of the single-cell sequencing, single-cell generally has a high noise sequence data, there is a high dropout problem that many genes are expressed or not expressed moderate detected efficiently. So, for the differences before the traditional multi-cellular transcriptome sequencing data to develop a method of detecting the expression or the software may not be entirely suitable for single-cell sequencing data. To differentially expressed when the cell, in order to obtain reliable results, necessary to select the different isoforms or a different cell conditions are good differential expression analysis method (micro-channel public number: AIPuFuBio).
In recent years, there are many differences expressed exclusively for a single cell transcriptome sequencing data analysis methods have been developed, such as MAST (Finak et al., 2015), SCDE (Kharchenko et al., 2014), DEsingle (Miao et al ., 2018), Census (Qiu et al., 2017), BCseq (Chen and Zheng, 2018) and the like. Specifically shown in the table below:
Above the red line is specialized expression analysis software packages or R, below the red for the differences in the development of single-cell sequencing data is for the bulk of transcriptome data development software packages or R
这里要值得提一下SCDE（全名：Single Cell Differential Expression）软件，其属于最早一批专门针对单细胞测序数据开发的差异表达分析软件，地址为： https://hms-dbmi.github.io/scde/。下图是原文章中SCDE与其他传统差异表达分析软件的性能比较，显示SCDE具有不错的性能。
图2、SCDE与其他软件在单细胞测序数据集上鉴定差异表达基因的性能比较（Kharchenko et al. Nature Methods, 2014）
最近，Wang et al.等人比较了11款经典的软件在单细胞测序测序数据集上的差异表达分析性能，这些软件具体如下表所示：
图5、不同差异表达软件各主要指标的比较（ Wang et al. BMC Bioinformatics, 2019）
图6、不同差异表达软件之间在真实数据集上检测到的差异表达基因比较（ Wang et al. BMC Bioinformatics, 2019）。差异表达基因的定义为：adjusted p-value< 0.05
图7、样本数量对不同差异表达软件各方面性能的影响比较（ Wang et al. BMC Bioinformatics, 2019）
图8、不同差异表达软件鉴定到的top 300个差异表达基因富集的显著KEGG通路数和GO条目数比较（ Wang et al. BMC Bioinformatics, 2019） 。（条件：FDR<0.05）
总的来说，不同的差异表达软件有不同的优缺点。有些软件具有高灵敏性，但检测精度却比较低，有些则刚好相反。这11款软件中，DEsingle 和SigEMD这两个方法较好的平衡了差异表达基因检测灵敏性和准确性。值得注意的是，Wang et al. 的比较发现，现在这些专门针对单细胞测序数据开发的差异表达分析软件和传统的方法相比，并没有显示出太多的优势（ Wang et al. BMC Bioinformatics, 2019）。这也进一步说明，还需不断开发新的单细胞测序差异表达分析方法，以更好的检测单细胞测序数据的差异表达基因。（更多经典，可见大型免费综合生物信息学资源和工具平台AIPuFu：www.aipufu.com）。笔者建议，做单细胞测序数据的差异表达分析，最好还是选择专门针对单细胞测序数据开发的软件，如SCDE、DEsingle 和SigEMD等。
希望今天的内容对大家有用哦，会持续更新的，欢迎留言~~
参考文献
1. Chen et al. Single-Cell RNA-Seq Technologies and Related Computational Data Analysis，Frontiers in Genetics, 2019
2. Wang et al. Comparative analysis of differential gene expression analysis tools for single-cell RNA sequencing data, BMC Bioinformatics, 2019
3. Kharchenko et al. Bayesian approach to single-cell differential expression analysis, Nature Methods, 2014