Minimum dsDNA detection by NGS machine

[cosmarium]

Hello,

I am new here and to NGS. I prepared some multiplex DNA libraries with Illumina TrueSeq Kit. I submitted each sample/index at 10nM, to the facility that operated the sequencer. They told me that they used 1uL from each sample to make a 12uL total volume of pooled samples for one lane. This means each library was at 0.8nM. I got enough reads for that, and they used a GAII sequencer.
Now I have prepared samples for 8 lanes, multiplexing, in the same fashion. For a reason, we have to use other facility with a HiSeq 2000 sequencer. They asked me to have each library at 10nM. I don't have that concentration. Each library is 1.09nM for the lane with the least concentrated libraries. Other lanes are similar, maximum value is 3.38nM/per library (index). For the lane of 1.09nM I have 12 samples/indexes, so the dsDNA concentration for that lane is 13.08nM.
I think I will get enough coverage (72 cycles), but I will like to know if this concentration of dsDNA is enough for the machine to detect and amplify.
Please help me, I am so afraid to lose the work I have done for months :/
I put a lot of effort in my work, this is getting frustrating.
Thanks
C

[Heisman]

Not quite sure I completely understand your question, but I think you're fine; they want your pooled sample to be at 10 nM, not each individual index, most likely.

[cosmarium]

Minimum dsDNA detection by NGS machine

Thanks for replying.

My question can probably restated as:
Would the HiSeq be able to form cluster with libraries at 1.06nM each?


Further info:
This is the comment from the person managing the machine that I am going to use.
"When pooling libraries, don't 1st dilute to 10 nM and then combine because when combining equal amounts of 10 nM solutions you will dilute the final concentration of each to 0.833 nM. Instead calculate how much of each library is needed to make a 10 nM solution in say 50 µl volume. Mix those together and dilute to 50 µl. This way the final concentration will be 10 nM for each library. In this way when we hybridize in our normal fashion, we will be able to get the maximum # of sequence tags from your lanes".
So, I think the 10nM concentration is for optimal amplification of libraries. I think I have enough material to get an amount of amplification that covers my needs, based on the fact that the machine can pick up my low? concentrated DNA.
I think this machines are very sensitive, for cancer research -and this is not the only example- researchers may just have very small amount of input DNA. So, I am ok I think.
But when I look of input DNA for Illumina sequencing all information that I find is about input amount for sample prep not for the actual cluster formation.

Thanks again,
C

[Heisman]

Well, that is new to me. Our sequencing core specifies to submit libraries at 10nm. The idea being that if we submit at 10nm and then it is loaded at an optimal concentration (say, 10pm), then we will get an optimal cluster density. And we use HiSeq 2000's. Whether each index is at 10nm or 1nm doesn't matter as long as they are all equal (if you want the same amount of reads for each index); it's the loading concentration of the entire library that determines the cluster density.

[pmiguel]

Add your samples into the pool you wish to use in the ratios you want, then concentrate your sample to 10 nM. I would recommend a speedvap -- preferably with no heat. Make sure to use a low bind tube and/or final 0.1% TWEEN. Also good to do a final pool concentration check to make sure nothing went awry.

The v3 cBot protocols call for 2 nM initial concentrations. Maybe your core would use the v3 protocol for you?

"When pooling libraries, don't 1st dilute to 10 nM and then combine because when combining equal amounts of 10 nM solutions you will dilute the final concentration of each to 0.833 nM. Instead calculate how much of each library is needed to make a 10 nM solution in say 50 µl volume. Mix those together and dilute to 50 µl. This way the final concentration will be 10 nM for each library. In this way when we hybridize in our normal fashion, we will be able to get the maximum # of sequence tags from your lanes".

Makes no sense unless they are adding a dilution factor for each library in the pool down stream. The way the v3 cBot protocol works is:

-- 10 ul of your 2 nM pool is mixed with an equal volume of 0.1 M NaOH. So now you have 20 ul of 1 nM library.

-- The 1 nM library is diluted to 1 mL with "HT1" buffer. A 1:50 dilution! Now your library is at 20 pM.

-- Dilute 20 pM library down to your preferred working range. This is usually in the 10-14 pM range. This gives you more than enough for an entire flow cell. The cBot wants to be loaded with 120 ul for each lane. It sips about 100 ul/lane during cluster generation.

Anyway, I think there is a miscommunication between you and your core.

To answer your intial question -- how "minimun dsDNA detection" on a HiSeq.

The minimum threshold is constrained by two main factors:

(1) The minimum cluster density below which the cluster calling software will fail to register a panel. I am not sure where this is exactly, but 300,000 clusters/mm2 is a safe lower boundary. If there are no problems with you library, that should be easy to exceed with a 10 pM solution of your library loaded onto the cBot.

(2) The maximum permitted NaOH concentration in the libraries loaded on the cBot should be no more than 1 mM. That means a 1:50 dilution is called for. Also, you want to mix your NaOH and sample in equal amounts. So that calls for another 1:1 dilution which gives you a 1:100 dilution all told.

So, in principle, one could start with 1 nM sample and still gather data from a run. But 2 nM is better, because then the sample can be adjusted into the ideal range for the sequencer.

In practice you want to give your core libraries in the concentration they ask -- if you trust them. The text quoted above looks like it would end with the flow cell being 6-12x over clustered. Which would be a disaster. But if your core is typically producing good results, then you might want to go to another lab that is submitting samples to them with good results and see what they submit.

--
Phillip

[cosmarium]

Thank you very much you all.

I dropped my samples today, the people at the core facility is going to analyze them by qPCR and bioanalyzer sizing chip. After this, they will tell me if the samples are ok to run.

The woman who prepares the samples asked me to give her the concentration of my samples in nmoles per uL, although I gave her concentrations in nM concentration...?

She also asked me at which working concentration I specifically want my samples (otherwise she can't prepare them). Because of your comments I was able to understand she was asking me where in the 10-14pM range I want my samples to be. How I know this?
Any help will be greatly appreciate it.
Thanks again in advance,

C

[cosmarium]

I work with E. coli.

[Heisman]

She probably meant nanomoles per liter, which would be nanomolar.

I would to 10pm if this is your first time to be safe... if it's underloaded you'll still get a decent amount of usable data; if it's overloaded you may get very little usable data.