Question About Direction of RNA-seq Data

[lylyf1987]

Hi,

I have a simple question.
For the single end reads, do they all come from the 5'end of the fragment or 3' or both? And does Fastq file include this information?

For the paired end reads, does one Fastq file save all 5'end reads and another save all 3' end reads? If so, could I know which file is for 5'end and which for 3'?

[dpryan]

This depends on the kit used. Some kits are directional, others aren't. For most directional kits, read #2 in a pair denotes the strand, which means for SE reads it's the opposite of how each read aligns. This is not encoded in any way in fastq files, though that's a great idea (particularly given how many headaches this sort of thing has cased).

For PE reads from a directional library, one file will contain reads mapping to the sense strand and the other will map to the anti-sense strand. Most protocols use the dUTP method, so read #2 will then map to the sense strand (unless I'm reversing that in my head!). When in doubt, just align everything and look at the results in IGV. You can color reads according to orientation and then just look at a couple genes to confirm which read is denoting orientation (since it'll match the direction of the gene). Of course, if you used a non-stranded kit then you'll just see read #1 and #2 each roughly equally balanced between both orientations.