RNA Sequencing of Formalin-Fixed, Paraffin-Embedded Specimens for Gene Expression Quantification and Data Mining
RNA Sequencing of Formalin-Fixed, Paraffin-Embedded Specimens for Gene Expression Quantification and Data Mining
复制标题
DOI：
10.1155/2016/9837310
复制
发表时间：
2016-01-01
影响因子：
2.9
通讯作者：
Cai, Qiuyin
中科院分区：
文献类型：
--
作者：
Guo, Yan;Wu, Jie;Cai, Qiuyin
背景在研究基因表达时，适当的rRNA消耗对于成功利用FFPE样本至关重要。我们进行了一项研究，以评估两种主要的rRNA去除方法：Ribo-Zero和RNase H。从4个样品中提取的RNA用两种rRNA去除方法一式两份处理并测序（N = 16）。我们评估了它们的还原能力、检测RNA的能力以及对这些三阴性乳腺癌标本进行分子亚型分析的能力。结果两种rRNA消除方法在技术重复之间产生一致的数据。我们发现，与Ribo-Zero方法相比，RNase H方法产生了更高质量的RNAseq数据。此外，我们评估了从FFPE组织样本中产生的非编码RNA的RNAseq数据，包括lncRNA、增强子/超级增强子RNA和单核苷酸变异（SNV）。我们发现，与Ribo-Zero相比，RNase H更适合检测高质量的非编码RNA，并在重复之间提供更一致的分子亚型鉴定。不幸的是，这两种方法都没有产生可靠的SNV数据。结论.总之，对于FFPE标本，RNase H rRNA耗竭方法的性能优于Ribo-Zero。两种方法都不能产生足够用于SNV检测的数据。
Background. Proper rRNA depletion is crucial for the successful utilization of FFPE specimens when studying gene expression. We performed a study to evaluate two major rRNA depletion methods: Ribo-Zero and RNase H. RNAs extracted from 4 samples were treated with the two rRNA depletion methods in duplicate and sequenced (N = 16). We evaluated their reducibility, ability to detect RNA, and ability to molecularly subtype these triple negative breast cancer specimens. Results. Both rRNA depletion methods produced consistent data between the technical replicates. We found that the RNase H method produced higher quality RNAseq data as compared to the Ribo-Zero method. In addition, we evaluated the RNAseq data generated from the FFPE tissue samples for noncoding RNA, including lncRNA, enhancer/super enhancer RNA, and single nucleotide variation ( SNV). We found that the RNase H is more suitable for detecting high-quality, noncoding RNAs as compared to the Ribo-Zero and provided more consistent molecular subtype identification between replicates. Unfortunately, neither method produced reliable SNV data. Conclusions. In conclusion, for FFPE specimens, the RNase H rRNA depletion method performed better than the Ribo-Zero. Neither method generates data sufficient for SNV detection.