What More Than Gene Expression Analysis Can I Do With My Rna-Seq Data?

Question

[jobinv]

My research group has performed this study where we have implanted a cell line derived from a melanoma brain metastasis in a patient, into a number of mice. These mice subsequently got new metastases to different organs, and we have taken samples from tumors in four different organs (brain, ovaries, adrenals, bone), three replicates from each location, and performed RNA-sequencing on them. We've also performed some exome sequencing.

Now, all that I've done so far is to attempt to find a gene signature for the metastases to the different organs, based on the gene expression. This is interesting in itself of course, but I am painfully aware that there is ever so much more that is possible to extract from the sequencing data that we have performed. The problem is that I'm very fresh in this game, and have only recently started learning how to do bioinformatics analyses. My question is whether anyone can advise me in how I can attack this data in more ways than what I've done so far? It would be much appreciated!

Answer 1

[k.nirmalraman]

Differential Gene Expression Analysis (obvious)

Network Analysis

Gene Set Enrichment Analysis

If your sequencing method permits, you may also analyse splice variants

There is still a debate, if you can detect SNPs using RNA Seq analysis (I personally dont know if this would be possible with the RNA Seq chemistry.. However, there are a few approaching it to this end here.

[jobinv]

Thank you, I will definitely check those out!

[Enx]

Also it is possible to look for fusion genes in RNA-seq data!

[Vitis]

Sure you can detect SNPs from mRNA-Seq data, although in very specific settings. I've identified EMS point mutations from mRNA-Seq data twice, but with extensive biological information, such as phenotype, genetics, etc.

Answer 2

[User 59]

If you have matched exome data then you could consider doing some allele specific expression work

[jobinv]

Would you mind elaborating on this? I'm not sure what this means, you see.

[Steve Lianoglou]

Assume that there are no CNVs for a particular gene, then you'll have two copies of this gene (on sister chromosomes). The suggestion is to see if there is differential expression of the gene between chromosomes.

[jobinv]

Oh, wow, that is interesting. Thanks!

Answer 3

[Enx]

Also one could look for large scale variations in RNA-seq, like for example fusion genes, translocations, etc. It is known that some fusions genes have key roles in some tumors, like for example TMPRSS2-ERG fusions in prostate cancer! Most of the fusion genes are found using RNA.-seq data. FusionCatcher http://code.google.com/p/fusioncatcher/ can be used for example for finding fusion genes and translocations in RNA-seq data. Here is more info about http://code.google.com/p/fusioncatcher/