Hello, I am currently thinking about analyzing the microbiome from mouse's gut. I will do the analysis on 454 sequencer. What I do know is that I will prepare the library using the paired end approach, but sincerely I don’t know how many sequencing reaction I should do for a sample? Will one large region of the PicoTeiterPlate be sufficient?
I'm afraid this is one of those questions for which there really isn't a universal answer. It very much depends on what question you plan to bring to the data.
I don't have much in-depth knowledge about the 454 platform (we tend to use Illumina exclusively) but i think the fundamental question about sequencing depth is independent of platform.
One thing that is not clear from your question if is you plan on using a full shotgun approach or only focusing on the 16S gene. If you do choose the latter, then you can actually get by with a shallow run. People report really high resolution runs in most papers, but then they rarefy to the lowest number among their samples (usually somewhere between 5000-20,000) reads, effectively throwing everything else away. This leads me to conclude that you can get a decent framing of the community with something as low as 5000 reads per sample. (I remember a publication going as low as 1000, but i can't seem to find it now).
If on the other hand you're going for the "full on" approach, you'll need a lot more depth to just recover the diversity that you would see from a shallow 16S run. In this case I would argue you should go for as deep as you can afford.
Thank you Paul. What I want to do is to make "full" approach. But I do not know if I can afford it. I can't plan one sequencing run for a sample, and then realize that it is not enough. Because during the evaluation of grant I will be"executed" :)
- Badges
- Science topic
- Similar topics
- Linguistics
- Philology
- Language
- English