Comparison of RNA-seq and microarray platforms for splice event detection using a cross-platform algorithm

Abstract

Background: RNA-seq is a reference technology for determining alternative splicing at genome-wide level. Exon arrays remain widely used for the analysis of gene expression, but show poor validation rate with regard to splicing events. Commercial arrays that include probes within exon junctions have been developed in order to overcome this problem. We compare the performance of RNA-seq (Illumina HiSeq) and junction arrays (Affymetrix Human Transcriptome array) for the analysis of transcript splicing events. Three different breast cancer cell lines were treated with CX-4945, a drug that severely affects splicing. To enable a direct comparison of the two platforms, we adapted EventPointer, an algorithm that detects and labels alternative splicing events using junction arrays, to work also on RNA-seq data. Common results and discrepancies between the technologies were validated and/or resolved by over 200 PCR experiments.

Results: As might be expected, RNA-seq appears superior in cases where the technologies disagree and is able to discover novel splicing events beyond the limitations of physical probe-sets. We observe a high degree of coherence between the two technologies, however, with correlation of EventPointer results over 0.90. Through decimation, the detection power of the junction arrays is equivalent to RNA-seq with up to 60 million reads.

Conclusions: Our results suggest, therefore, that exon-junction arrays are a viable alternative to RNA-seq for detection of alternative splicing events when focusing on well-described transcriptional regions.

Keywords: Alternative splicing; Microarrays; RNA-seq.

Fig. 1

EventPointer overview for junction arrays and RNA-Seq data. a The CEL or BAM files are the input data for each technology. The splicing graph for each gene is built using the array annotation files or directly using the sequenced reads. b Each node in the splicing graph is splitted into two nodes that correspond to the start and end positions in the genome respectively. EventPointer identifies events within each gene and annotates the type of event. In the figure, among the events in the gene, an exon cassette is highlighted. c Statistical significance of the events is computed. d Finally, the top-ranked events are validated using PCR and the results visualized in IGV

Fig. 2

Correspondence between the events detected by arrays and RNA-seq. An event is considered to be significant if the p.value is smaller than 0.001 and non-significant is it is larger than 0.2. Events with p-values between both are considered to be inconclusive cases

Fig. 3

FDR for different types of events using both technologies. Panel a shows the FDR for matched events. Panel b shows FDR for the events detected in each technology regardless or being matched or not. In both technologies, alternative 3′, 5′, first and last exons have larger FDR than other types of events

Fig. 4

Estimated PSI (for RNA-seq, microarrays and PCR image analysis), PCR bands, the reference HTA transcriptome and the alternative paths of the DONSON (panel a) and MELK (panel b) genes in R⁺M⁺. Each of the points represents the same replicate in either of the three technologies. The last numbers shown are expected bands for the selected primers. If the number is shown to the left side of the double bars, the band corresponds to Path 1 of the event (long path). If shown to the right side, corresponds to Path 2 (short path)

Fig. 5

Increment of PSI for both microarrays and RNA-seq. The black (gray) dots represent events with high (low) standard deviation in the differential usage of the isoforms in both paths. Correlation between events with high and low variability are 0.90 and 0.61 respectively

Fig. 6

a Events detected using RNA-seq and array technologies. b Type of event after matching the events detected by both technologies