MiSeq Error: "sequence contain no element"

[rnaeye]

Hi,
I got following error after an overnight run on myseq. I was wondering if anyone has any idea what went wrong.
Error message was "sequence contains no element".
Last time I checked the instrument it was at cycle-11 and everything was fine.

My first reaction was to call technical support, however they close due to time difference.

I will appreciate any input. Thank you.

[KymJJ]

Hi,

I was just about to log a post for the same issue when I saw your notice. I had the same problem, overnight run, got to cycle 90 ish on run 1 and stopped. Have spoken to tech support and they have no idea what happened, except that my error rate went through the roof. Tried running again tonight because we thought we had pinned it down to an interruption in the run, but it's done the same again - error rate through the roof and stopped around cycle 80.

Did your run complete or stop part way through? Are you on MiSeq version 1 or 2? What kind of libraries are you running?

[rnaeye]

Hi,

I was just about to log a post for the same issue when I saw your notice. I had the same problem, overnight run, got to cycle 90 ish on run 1 and stopped. Have spoken to tech support and they have no idea what happened, except that my error rate went through the roof. Tried running again tonight because we thought we had pinned it down to an interruption in the run, but it's done the same again - error rate through the roof and stopped around cycle 80.

Did your run complete or stop part way through? Are you on MiSeq version 1 or 2? What kind of libraries are you running?

I have V2. First time I was running 50 cycle sequencing. It stopped at cycle 37. Cluster density was 1400K/mm2. Tech support was very helpful, suggested that too high density clusters might a problem. I diluted my sample two fold, and tried to run again. As soon as I clicked on the run button, it gave me the same error. However, instrument continued generating clusters, and next day surprisingly I found that sequencing was successfully done. At my second run, my cluster density was 1350K/mm2, still higher than recommended density.

I went and double checked my DNA concentration. It was 4 fold higher than recommended!!! Looks like I made a careless math mistake during DNA dilutions.

After I brought down the DNA concentrations to recommended concentration (840K and 1000K/mm2 cluster density), I had two successful Miseq run. And today I am starting another one.

In short, my problem was the high cluster density. It's strange that instrument gives an error when cluster density is too high, and Myseq manual does not mention anything about this. I can understand that you need a certain cluster density for good sequencing; I would think that wrong density only would affect data quality. But, it looks like it causes an instrument error.

Please let me know how your problem evolves. I would be very much interested in knowing about it.

Good luck!

Thank you.

[KymJJ]

Thanks for getting back to me. Looks like it gives the same error for different issues - my cluster density was at about 800K so I think we probably have different things going on! Good to know you've got a solution though, and certainly handy to know for future. Tech support are onto my case now - I'll let you know if I get to the bottom of it!

Happy sequencing!

[yaximik]

I had MiSeq stopped in the middle of run at density of >1300, but with different error message, which however was similar in making a little sense, something "A reference object has no instance of an object". My tech support contact explained that such a message in computer language means the software cannot perform a task what it is directed to perform. My understanding is the actual message depends on a subroutine that eventually fails due to high cluster density. He said MiSeq will try to get data even if forced to operate a way outside the optimal range until something fails. In my case the signal intensity chart was very indicative in its weirdness, going over maximum and then dropping before the run was stopped.

[pmiguel]

All,
There is no way that high cluster density should cause the instrument to error our in this fashion. Basically it would be Illumina admitting that they did not allocate sufficient resources to their processing regimen.

At 4x the recommended cluster density it is unlikely you will get anything useful. So it doesn't really matter that the software bombed out. That is if your cluster density is really 3200 K/mm^2 all the clusters will pretty much merge into a big mess. The software might report 1300 K/mm^2, but that just means that it failed to find most of the clusters. So no big loss there.

However, if you had a true 1300 K/mm ^2 raw cluster density and the instrument errored out, that is a big problem -- it should be able to handle that under v2.

--
Phillip