[mdalessio]

I am a first time MiSeq user, so I am a bit confused. My understanding was that the MiSeq does its own clustering once you load the single sample well. Am I wrong?

[FredTam1]

I've been thinking that Qubit isn't as accurate due to similarities in it's reading of de-natured single strand vs. normal double-stranded DNA

[mcnelson.phd]

I am a first time MiSeq user, so I am a bit confused. My understanding was that the MiSeq does its own clustering once you load the single sample well. Am I wrong?

For the MiSeq, the clustering process itself occurs solely onboard the system. This is different from the HiSeq which used to require a separate piece of equipment, the cBot, to perform the clustering step on the flowcell which then had to be physically transferred to the HiSeq itself. For upgraded HiSeqs (1500/2500), the operator has the option of doing onboard cluster generation which allows for the longer read lengths but with the compromise of lower cluster densities and thus less data yield compared to using the cBot.

In both cases though, the user is responsible for ensuring that the correct concentration of library fragments are being added to the flowcell in order to build the clusters. For both systems, if you under estimate your library concentration and add too many viable fragments, then you'll form clusters that are too densely packed for the system to accurately determine the correct base call (which is determined by the chastity filter) and thus the number of clusters that pass filter will be very low. If it's too overly clustered, then the run may just fail because the software can't figure out signal from background.

Conversely, if you over estimated your library concentration then too few fragments would form clusters compared to the optimal cluster density. In this case, almost all reads will pass the chastity filter, but you'll be wasting reagent and system capacity. In the absolute worst case of this, you'll have so few clusters that the run will fail immediately after cycle 1 when the system is performing its focus imaging and determines there's not enough signal to differentiate clusters from background.

[dfhdfh]

We found the bead normalisation was OK but relying on it to achieve the correct cluster density was a wate of time. Most users want 1 good MiSeq run and the QT is too falky to guarantee this.
We real-time PCR QT teh final pool.
The most robust way is to use qPCR on all samplesand normalise before clustering.

Such a shame as the bead normalisation holds so much promise.

What primers do you use for the qPCR? So far I've only found the adapter sequences of the Nextera kit but not of the Nextera XT kit as provided by the Illumina sequence letter. Are the sequences the same for both kits?