MiSeq v3 2x300 run, Low PF and Q30

[SarahAurora]

I recently ran a ~200-400bp library prepped with NexteraXT on the MiSeq using reagents kit v3 for a 2x300 run, paired end, dual index reads: 10pM pooled library with 15% 10pM phiX.

I got a high cluster density of 1185 k/mm2 with only 27% clusters PF and >=Q30 14%.

Does anyone have any ideas for troubleshooting? Could this come from old kits or bubbles on the flowcell, would this even be visible? I did a qPCR (KAPA library quantification) before dilution all samples to 4nM. The qPCR worked fine, so there must be an error afterwards or with the kit.
Could it be that there was a DNase destroying the amplicons?

I'm curious about your ideas and answers.

Thanks in advance
Sarah

[GenoMax]

Have you checked the actual insert size of your library? It is possible that you have inserts significantly shorter than you expect them to be. That can explain the precipitous drop in Q score on R2.

You can check post #2 in this thread to see how you can calculate that information.

This run can also be borderline overloaded. That cluster density number is not very accurate once you get next the upper spec limit. Have you had Illumina tech support look at this run?

[SarahAurora]

Thank you for your reply. We have done the sequencing of this amplicon multiplex for many times and know the actual sizes. But are the moment the demultiplexing is problematic because it's too less reads.
The run could be overloaded but we have also done the sequencing with equal samples and even more input. It was very overloaded but we did get a lot of clusters passing filter (but a bad Q30). Now we have a very low clusters passing filter. So I think that overloading could not be the main issue.
No, we have not (yet) been in touch with tech support. I'm still hoping that the clues here could lead to an explanation.

[GenoMax]

If you have done similar sequencing successfully in the past then I suggest that you contact Illumina tech support and have them look at the run. This may turn out to be a hardware related problem on the sequencer.

[thermophile]

have you looked at the thumbnails? I'd guess your cluster density is higher than reported.

[SarahAurora]

Thumbnails of the tiles look ok, or let's say, they don't look as bad as the very overclustered run before. But they don' look balanced.

[GW_OK]

Can you post the %base graph?

[SarahAurora]

This one? I am not sure which one you want

[SarahAurora]

It's strange that we can read the first bases of R1, but no R4 and just a few Indices (R2+R3). Looks like the sequencing could not get read out like something intereferred with the optics.

I am very concerned about the run and my mind starts to melt, so please excuse this question: If I would have had trouble with the Index-PCR, then the adapters never would have gotten fixed on the flowcell, right? I want to exclude errors while preparing the run.

[GW_OK]

That's intensity. I was talking about the graph that is from % Base, which is selectable in the menu from the gray button under Intensity.

Still, though, in looking at this graph it appears you don't have a short insert but you do have sort've an odd 3-steps down to 0 intensity.

It looks to me like a hardware issue.

[SarahAurora]

Ok, thank you for your opinion. The longer I check all circumstances, the more it gets clear that the problem was not the DNA. Because at first the bases looked fine.

[GW_OK]

What's troubling to me is that you don't have any intensity in read 2, which should always show if you have at least a read 1, regardless of index.

Have you run any of the onboard diagnostic tests?

[GenoMax]

@Sarah: Do you not have a maintenance contract on this MiSeq with Illumina? If you do then you should go ahead and let tech support have a look at this run.

[SarahAurora]

What's troubling to me is that you don't have any intensity in read 2, which should always show if you have at least a read 1, regardless of index.

Have you run any of the onboard diagnostic tests?

Our bioinformatician told me we have 300 bp read 1 and read 2 but he main part is not Q30 or even Q20. So probably the DNA-bases part worked finde but the optics couldn't recognize anything...

No I did not know that those tests even exist...

[SarahAurora]

@Sarah: Do you not have a maintenance contract on this MiSeq with Illumina? If you do then you should go ahead and let tech support have a look at this run.

It's not our MiSeq, we just use it from our cooperation partner. I'll check if they have a maintenance contract with Illumina. Nevertheless, I wrote tech support an email with all details.