IGB Gallery
Tiling array example
Splicing and ESTs example
Tiling array expression data and pyrosequencing ESTs from light-grown seedlings
The image below presents shows a side-by-side comparison of expression tiling array data (bottom) and EST coverage (top) from light-grown seedlings.
The tiling array data are from Yamada et al, and the ESTs are from Weber, et al. (2007). ESTs were generated from a single run of a GS 454 pyrosequencer. Genomic alignments of the ESTs were made using blat. We obtained the alignments in PSL format from the authors. To simplify the display, we retained alignments for ESTs that mapped to one location in the genome.
To create the view of EST data shown in the image, we opened the filtered, blat output PSL file (using File->Open), which created forward and reverse strand tracks (also called tiers) containing alignments of the 454 sequences. We then created an annotation depth track from the ESTs, a genome graph that displays the number of overlapping features at every base pair position. To do this, we right-clicked the tier label (at left) for the 454 EST track and selected the "Make Annotation Depth Graph" option.
Expression array and EST coverage
Conclusion: The EST and tiling array data were generated in two different labs, at different times, but, in general, the data are highly correlated, at least with respect to the overall level and pattern of expression of known genes in the Arabidopsis genome.
Visualizing alternative splicing - finding the dominant splicing form using ESTs
These two images show views of two loci annotated by TAIR as being alternatively-spliced. In one case (AT1G32410), differential splicing affects the five-prime untranslated region and there are several overlapping ESTs that support both forms, although one form clearly predominates. In another case, alternative splicing affects the coding region, resulting in two distinct protein products. Comparison with overlapping ESTs gives a rough estimate of the relative prevalence of the alternatively spliced forms.
ESTs support two different forms of an alternatively-spliced exon.
One EST (selected) supports the shorter version of an alternatively-spliced exon in the protein-coding region of the transcript.
To create these views, we launched IGB and loaded the entire mRNA data set, corresponding to the TAIR gene models. We scanned the genome, looking for locations where multiple gene models were stacked on top of each other, a visual cue that the TAIR gene models in the region exhibit alternative splicing or alternative promoters. We zoomed in on these regions and used the Data Access tab and the "Refresh Data" to load EST alignments overlapping the region from the Bioviz server.
The EST data set was created as follows: At the time the image was taken, there were around 1.5 million Arabidopsis thaliana ESTs in GenBank. We downloaded their sequences from GenBank, used mRNA/EST genome alignment tool blat to align them to the Arabidopsis genomic sequence, and then performed a quality control filter following guidelines published on the UCSC Genome Web site. (We used the same filtering method described for the UCSC human ESTs data set.)