RPKM Values and Cufflinks

[bonifera]

Hi,

I am in the process of identifying differentially expressed genes in Arabidopsis and want to filter out mitochondrial DNA and plastid DNA.

I ran TopHat using the iGenome from Illumina and ensembl. I then used cufflinks to quantify the abundance of RNA. To filter out the unwanted alignments, I used the "-M" option of cufflinks and included a GTF file containing all of the Mt genes and Pt genes. My results were just the same as before, but without the excluded genes.

Long story short, the RPKM values did not change at all. Using this option simply filtered the genes out. I guess I was expecting my RPKM values to change.

From cufflinks documentation with regards to the -M option:

"Tells Cufflinks to ignore all reads that could have come from transcripts in this GTF file. We recommend including any annotated rRNA, mitochondrial transcripts other abundant transcripts you wish to ignore in your analysis in this file. Due to variable efficiency of mRNA enrichment methods and rRNA depletion kits, masking these transcripts often improves the overall robustness of transcript abundance estimates. "

In order for me to actually obtain adjusted RPKM values, do I need to rerun tophat using a reference set that does not have these genes? If so, would that just include deleting portions of the fasta genome sequence and the GTF reference? Or is there another way I can work this issue?

Thanks!

[dpryan]

You don't need to rerun tophat2. All you need to do is give cufflinks a mask file listing where the rRNAs are (if you know). Alternatively, make sure in cuffdiff to not use the classic FPKM normalization, since that's not robust (the DESeq-style normalization that I think cuffdiff does by default now is much more robust to this sort of thing).