Underclustering in MiSeq run

[Jellyfish]

Hi all, this is my first post here.

I set up a MiSeq run on Monday and I'm trying to troubleshoot why I have underclustering.

So I made a silly quantification mistake where I missed a step in my qPCR analysis, which basically resulted in me calculating my normalisation based off values what were half of what they actually were (e.g. I normalised sample 1 from '40nM' to 4nM, when in reality the original concentration was 80nM.)

..It was just a stupid mistake on my part, but this has obviously resulted in my libraries being twice as concentrated (pooled and normalised to 8nM rather than 4nM).

I diluted the '4nM' library to '15pM' and spiked in 1% phiX for my run....

So I apparently accidentally ran a 30pM library on the MiSeq. But the weirdest thing is that I actually have underclustering where I would have expected it to be massively overclustered! My cluster density is 729K/mm^2 and I can't work out why.

For reference: this is whole genome sequencing prepped with the TruSeq Nano LT kit and run with the 2x300 v3 reagents. I believe I should be aiming for a cluster density of 1200-1400K/mm^2.

Previous runs (with the 2x75 v3 reagents) where I used 15pM libraries have clustered at ~1100K/mm^2.

Any insight into this would be much appreciated! I have another run starting tomorrow and a third next week and I'd really like to get to the bottom of this!

Thank you!!

[GenoMax]

Have you looked at the cluster images? Do they seem to match the cluster number reported?

[Jellyfish]

I've attached a couple of the images - to me they look to be more overclustered than underclustered.

[ATϟGC]

You might want to compare the clustering at the top (where sample was loaded) versus the bottom.

With typical cluster generation there are more clusters at the top than the bottom of the lanes since DNA is bound as it flows over the cell and there is a decreasing amount of DNA library available for binding.

With over-clustering there are fewer clusters interpreted by the machine at the top since there are so many clusters close together and they are erroneously interpreted as fewer clusters at the top compared to the bottom of the cell.

[Jellyfish]

Clustering at the top and bottom look very similar to me.

[ATϟGC]

Is that judged from the thumbnail images or the densities as measure in the flow cell chart of SAV?

[Jellyfish]

So the thumbnails (I've only looked at a few though!) look the same.

The flow cell chart shows higher clustering at the beginning and lowered at the end, as expected.

Also BaseSpace hasn't updated itself for the whole of read 2, so I'm a little concerned about that as well! The MiSeq looks happy however so hopefully that's just a basespace issue...

[Genetic Librarian]

The reported cluster density is not always reliable.
If a run is massively overclustered, the MiSeq can not distinguish single clusters and logically can not count them reliably.
Another indicator is the PhiX Alignment. If you spiked in 1%, but your SAV only reports 0,2% aligned, you might have overloaded your library 5-fold.

https://support.illumina.com/content/dam/illumina-marketing/documents/products/other/miseq-overclustering-primer-770-2014-038.pdf