RNA-Seq has been recently developed for transcriptome studies, including the discovery and quantification of genes, exons and isoforms, alternative splicing and differential expression. We participated in the MAQC-III/SEQC consortium led by FDA (Leming Shi, PI), to conduct a dedicated multi-site multi-platform experiment with several built-in ground truths to evaluate the performance of this technology. The consortium systematically analyzed the data to establish best practice protocols on quality control, mapping pipelines, data analysis pipelines and external RNA controls (ERCC) spike-ins.
Our results show that RNA-Seq is comparable to microarrays for differential expression profiling. At the gene level, RNA-Seq match to microarray in terms of reproducibility, accuracy and detection power, when sequenced at a depth higher than one fourth of a HiSeq2000 lane per sample, or 90M total reads. In addition, RNA-Seq has the power to make new discoveries, with sufficient depth.
Fold change correlation
False Positive Rates
We plan to further compare RNA-Seq with an updated version of GG-H which has many more probes per gene/exons/junctions than the commercial arrays.
The goal of the MAQC-III/SEQC consortium isto assess the technical performance of next-generation sequencing platforms by generating benchmark datasets with reference samples and to evaluate advantages and limitations of various bioinformatics strategies in RNA and DNA analyses.
The conclusions and recommendations from the study should be useful for regulatory agencies, study committees and independent investigators that evaluate methods for global gene expression analysis.
1. Rigorously assessed the performance of RNA-Seq in terms of accuracy, reproducibility, detection power and ability to make new discoveries, and compared to that of microarrays.
2. Created a well-designed experimental data set of RNA-Seq with unprecedented sequencing depth, and associated protocols.
1. Assess RNA-Seq for clinical samples, especially FFPE tissues.
2. Compare RNA-Seq with an updated version of GG-H for clinical samples.
Weihong Xu, Wenzhong Xiao, the MAQC-III/SEQC Consortium