paired or SE are sufficent

[mathew]

I have a data-set for RNA seq which somehow the lab has sequenced with Single end reads (I would have suggested PE). Originally plan was to use it for differential expression of human genes, transcripts. But later on I learned that the cell which was used was from a naturally infected with virus. My question is will it be reasonable if I align with viral genome and try to find out differential viral genes affected or it may be best to have PE for more specificity/ have more data to be more confident about viral genes.

[mathew]

Viral genome

Any suggestion please

[arkal]

I have a data-set for RNA seq which somehow the lab has sequenced with Single end reads (I would have suggested PE). Originally plan was to use it for differential expression of human genes, transcripts. But later on I learned that the cell which was used was from a naturally infected with virus. My question is will it be reasonable if I align with viral genome and try to find out differential viral genes affected or it may be best to have PE for more specificity/ have more data to be more confident about viral genes.

So i tried somethign similar to what you're doing once... i aligned human reads to a viral genome and even with paired end i got a lot (quantitatively, not relatively... i.e numerically a lot but relatively was ~0.01% of the human fastq reads file) of non-singleton maps (possibly due to certain homologous regions).... so if ur question is to figure out the viral genes contamination on your organism, i think u might need a different more specific approach.

[mathew]

Viral genes paired or single end

Thanks for your reply. Actually, my aim is not to find out virus as conatminant, but study both viral as well as human genes in a cellline which is naturally infected with virus. I thought once we align with himna genome- say we have around 80% reads and then with virus even though geneome is of small size with single reads we may not have enough to detect viral genes, I ahve alos been searching litearture and find- http://jvi.asm.org/content/86/3/1458.full where they have used PEX 100.

Any suggestion I am still debating how to procede?

Thanks

[arkal]

Hmmm... you can try what they did in that paper.... align with strict criteria i guess to get better results... i'm not sure about RNA seq aligning but as i said, when i aligned DNA reads from a non infected human genome to a viral genome i got some reads aligning (i never went further to check why) so you might need some sort of filter to remove these anomalies (true negatives i think, no?)!