Systematic comparison of RNA-seq pipelines for absolute and relative gene expression quantification

At present, it is admitted that RNA-seq is a more powerful and adaptable technique than hybridization arrays. Nevertheless, as RNA-seq needs a more complex data analysis, it has generated a lot of research on algorithms and workflows. This has resulted in an exponential increase of the options at each step of the analysis. Consequently, there is no clear consensus on the appropriate algorithms and pipelines that should be used to analyse RNA-seq data. In the present study, 192 pipelines on 18 samples from 2 human cell lines were evaluated. Absolute gene expression quantification was assessed by non-parametric statistics to measure precision and accuracy. Relative gene expression performance was estimated testing 19 differential expression methods. These results were contrasted in parallel with the microarray HTA 2.0 data from Affymetrix using the same set of samples. All procedures were validated by qRT-PCR on 32 genes in all samples. In addition, this study proposes a new statistical approach for precision and accuracy evaluation on real RNA-seq data. It also weights up the advantages and disadvantages of the algorithms and pipelines tested and gives a guide to select the appropriate pipeline to analyse RNA-seq and microarray data. Poly A+ RNA from KMS12-BM and JJN3 cells untreated or treated with amiloride or TG003 (0.1 mM, 0.4 mM and 0.4 mM, respectively) for 24 h was isolated and prepared for microarray hibridization.

目前学界普遍认可，RNA测序（RNA-seq）是一种相较于杂交芯片（hybridization arrays）更具效能与适应性的技术。然而，由于RNA-seq所需的数据分析流程更为复杂，相关算法与工作流的研究大量涌现，使得分析各环节的可选方案呈指数级增长。因此，目前学界尚未就分析RNA-seq数据的适配算法与分析流程达成明确共识。本研究针对源自2株人类细胞系的18个样本，共评估了192套分析流程。采用非参数统计方法开展绝对基因表达定量分析，以评估其精准度与准确度；通过测试19种差异表达分析方法，估算相对基因表达的性能表现。同时，本研究将上述结果与Affymetrix公司的HTA 2.0芯片数据（采用相同样本集）进行平行对照分析。所有分析流程均通过实时荧光定量反转录PCR（qRT-PCR）对所有样本中的32个基因完成验证。此外，本研究提出了一种全新的统计学方法，用于真实RNA-seq数据的精准度与准确度评估。本研究还权衡了所测试的各类算法与分析流程的优劣，并为选择适用于RNA-seq及芯片数据分析的合适流程提供了指导性方案。将未接受处理，或经浓度为0.1 mM的阿米洛利（amiloride）、浓度分别为0.4 mM与0.4 mM的TG003处理24小时的KMS12-BM与JJN3细胞中的聚腺苷酸化阳性RNA（Poly A+ RNA）分离并制备后，用于芯片杂交实验。