Transcriptome sequencing detects differentially expressed transcript variants. Because of the diverse nature of the MAQC A and B samples it is expected that there may be many genes with differentially expressed transcript variants. For example the original MAQC study found that the ELAVL1 gene was more highly expressed in Sample B than A on microarrays with probes designed to the 3' end of the gene, but more highly expressed in sample A on microarrays and the TaqMan, StaRT-PCR, and Quantigene with probes in the 5' end. Using the transcriptome sequences from both the ODT and TSEQ preparations, the differential expression of the two transcript variants in the A and B samples resulting from alternative polyadenylation sites is immediately clear. (a) The BLAST start and end positions of each read are plotted on the x and y for every read that hit the ELAVL1 gene in Samples A (red points) and B (blue points). Plotted this way, the forward reads correspond to the points above the gray 45 degree line and the reverse reads below. The reads for the brain Sample B cover the entire 6075 bp length of the RefSeq gene from the 5' to 3' end while most the reads for the pooled cell line Sample A hit a much shorter transcript with a truncated 3' UTR. This conclusion is also confirmed by the Affymetrix Exon 1.0 array. (b) The hybridization results for the core exon probesets for the ELAVL1 gene are plotted for Samples A (red points) and B (blue points) showing the differential expression of the short and long transcript variants. The black and grey bars at the top of (b) show the coordinates of the 6 exons of ELAVL1. The red diamonds indicate an alternative polyadenylation site in the middle of the 6th exon identified by AceView.
Mane et al. BMC Genomics 2009 10:264 doi:10.1186/1471-2164-10-264