P035
Comparisons Between RNA-Seq and Microarrays are Dependent upon the Depth at which you Sequence, Transcriptome Size, and Resolution of Expression Events you are Targeting
Identifying expression changes has evolved from simple gene expression experiments to identifying non-coding, miRNA and alternative splicing events. Thanks, in part to RNA-Sequencing, investigators today are more aware of these complexities and can obtain a higher resolution of biological change when applying new methodologies to their study designs. However, there are limitations to these available technologies. Here we will contrast the advantages and disadvantages if RNA-seq and microarrays and highlight new ways of visualizing complex expression results.
Results
The same RNA samples were applied to both RNA-seq and Whole Transcriptome microarrays and evaluated for precision and accuracy. The sequencing data was generated as part of the third phase of the MAQC project (MAQC-III), also called Sequencing Quality Control (SEQC). Comparison results between microarrays and RNA-Seq are completely dependent upon the depth at which the samples are sequenced. For RNA-seq there is an inverse relationship between precision and expression abundance. However for microarrays precision remains constant throughout the dynamic range. In terms of accuracy based on a tissue mixture model, two full lanes of sequencing per sample are required to get equivalent power seen with a whole transcriptome microarray.
Conclusion
Microarrays are comparable to RNA-Seq for differential expression profiling. However our data show that sequencing depth is a primary consideration to get equivalent results from a microarray. Medium and low abundance transcripts require deeper sequencing in order to get precise measurements. The depth required also depends on the species and cell types being explored as transcriptome sizes vary.