how much DNA do I need for paired-end sequencing?

[Simone78]

Hello everybody,
I am doing library prep for paired end BISULFITE seq on Illumina GAII. I am a little bit confused about the amount of DNA that is required for sequencing. They say 8 pM. After adaptor ligation and size selection I have fragments around 400 bp.
I found this formula to convert from pM to ng:
ug = (pmol * N * 660)/ 10^6

where:
pmol=8
N=400

but this corresponds to 2.1 MICROGRAMS of DNA!

After size selection of the 400 bp fraction I performed bisulfite treatment and then PCR. I should have a geometric increase of my DNA after each cycle, but my DNA is (must be) damaged from the harsh conditions of the bisulfite reaction, because I don't get as much as expected after PCR.
That means I need to go up with the number of PCR cycles (which I don't want), but even then I'm not sure I can reach 2 ug!
Am I wrong in my calculation? Any experience with the same type of experiment?
Thanx a lot!

[csc]

Just a quick observation. pM is a concentration. pmol is an amount. Are you sure you have the units correct?

[Simone78]

yes, you are right but in a paper from Lyko et al. (The honey bee epigenomes, PLoS Biology , Nov 2010) they say 8 pM, which seems strange to me.
I'm using exactly the same protocol and my libraries have conc in line with what is reported on that paper (0.8-5.8 ng/ul), but this seems very little.

They should give the AMOUNT of DNA they used and not a conc. So I thought it was a typo and they meant 8 PICOMOLES. Otherwise, how much of this 8 pM solution did they use? They don't mention anything else. Sorry, but still I don't get it...

[csc]

Ok, I went and looked up the reference. I think I understand the confusion. Ideally, in some sort of assay, one would present the absolute amount of material used (or concentration given a fixed volume). In Illumina sequencing-speak, the way they phrased it makes sense to me. What they are saying is they introduced a diluted library at 8pM into the flow cell in each lane (8pM is generally the concentration used to give optimal cluster density on a GAII). The sequencing protocol would tell you that you need 120uL at this concentration per lane. However, 8pM is after hydroxide denaturation and dilution into hybridization buffer, you will most likely want your final library at a concentration of >1nM (ideally 10 nM) but you will only need several microliters at this concentration. Make sense?

[Simone78]

Thank you very much! Now it is clear...and it turned out I got enough DNA after PCR!

[kmpphd]

(660 pg/1 pmol of bp)(8 pmol of dsDNA/1 L)(1 L/10^6 uL)(120 uL of dsDNA*/lane)(400 bp/1 dsDNA)(1 ng/10^3 pg) = 0.253 ng dsDNA/lane, which converts to 1.2x10^9 ssDNA molecules per lane.

*for calculation purposes

[TonyBrooks]

I would double check with who's running the sequencing what they load onto the flowcell too. We found that 8pM is fine for PhiX, but, from our experience, our own libraries have to run at almost twice that concentration to achieve a decent amount of clusters.

Also, before loading onto a flowcell, best practices indicate you need to make a 10nM stock, then denature in a 2nM NaOH reaction, You can skip the 10nM stock step and go straight to the 2nM denaturation if your DNA is limiting. There are also protocols on other threads here about denaturing really low concentration libraries, but the lowest I've seen done is a denaturation at 200pM.