IsoProfiler now provides access to GTEx tissue expression data, enabling you to seamlessly explore your human expression data in the context of RNA-seq expression data from 51 different human tissues. This enables you to answer questions like:
Is this differentially expressed isoform in my dataset specifically expressed in a particular tissue or defined set of tissues? For example, is it potentially a target for drug intervention in a particular tissue?
Are there other isoforms for this gene also known to be expressed in that tissue?
What are all the liver-enriched (or other tissues) isoforms in my dataset?
Which isoforms from my dataset are tissue-enriched? (i.e. enriched in any single tissue or small set of tissues).
For a given gene in my dataset, which are the highest or lowest expressed isoforms in particular tissue(s)?
The underlying GTEx data was processed from the raw FASTQ files by OmicSoft (a QIAGEN company) and further processed by the IPA team to calculate median FPKM values for each transcript in each tissue in both the Ensembl 88 and RefSeq 80 gene models. This data was then integrated into IsoProfiler, along with a calculated tissue enrichment score. IsoProfiler enables you to filter the transcripts in your RNA-seq dataset based on whether transcripts are tissue-enriched by this definition (see the IsoProfiler help article for further detail).
Figure 1 shows a portion of an IsoProfiler window filtered to show the set of genes in a hepatocellular carcinoma RNA-seq dataset where at least one of its transcripts is 1) in the dataset 2) enriched in liver, 3) is protein-coding, and 4) where the gene is associated with hepatocellular carcinoma (based on the Ingenuity Knowledge Base):
Fig 1. IsoProfiler with new GTEx human tissue expression data. The filtered view above shows a subset of genes from a hepatocellular carcinoma dataset, which can be found in the IPA Example Analysis datasets folder (entitled “HCC EM pool Tumor vs_ Normal 2016-09-30”). The IsoProfiler filters were configured to retain all genes in the dataset which have at least one transcript enriched in liver (based on the GTEx data), coding for a protein, and where the gene is associated with hepatocellular carcinoma according to the Ingenuity Knowledge Base. The expression in liver (from GTEx) is indicated in the column “Liver (FPKM)” in the screenshot, and FPKM values that are considered liver-enriched are highlighted with bold text and a yellow highlight. See IsoProfiler help for details on IPA’s definition of tissue enrichment and how to use the new feature.
You can quickly visualize the expression of all of the isoforms for a given gene in the form of a line chart as shown below in Figure 2. Clicking on the “3 tissues” link (shown in Figure 1) for FGL1-204 brings up the chart with that isoform’s expression (from GTEx data) highlighted in yellow. It is highly enriched in the liver. Indeed, it is expressed over 9000 times higher in the liver relative to the median across the 51 tissues.
Fig 2. GTEx human tissue expression data for the gene FGL1. The line chart in IsoProfiler for the FGL1 gene indicates that the FGL1-204 isoform is expressed at a much higher level in liver than any other tissue (of the 51 available tissues), and that it is expressed much higher than other FGL1 isoforms. The chart represents the median FPKM for each transcript across the set of samples for each tissue listed on the X-axis.
This result of this combination of filters suggests that FGL1 and (more specifically the FGL1-204 transcript) could be a relevant biomarker: it is associated with HCC (from a curated paper in the Ingenuity Knowledge Base, PMID 14981537), encodes a protein, and is highly specific to liver. IsoProfiler with GTEx is a unique tool to enable you to conduct innovative research at the transcript level with RNA-seq data.
Note: IsoProfiler is only available when you have an Advanced Analytics license in IPA.