[james hadfield]

Tom - the MiSeq moves your sample from the cartridge to teh flowcell through non-disposable tubing, there is a risk some carry-over will happen but this can be mitigated against with several relatively small changes.

Many applications will be largely unaffected, amplicons in heterogenous samples are likely to suffer the most.
Send your samples with different barcodes each time if you can.
Make sure to ask your facility to perform the washes between runs.

There is a hack to fool the MiSeq into thinking it is clean, don't use it! A maintensance wash is a bit of a pain but we'll be running them until this is fixed.

[thomasblomquist]

Tom - the MiSeq moves your sample from the cartridge to teh flowcell through non-disposable tubing, there is a risk some carry-over will happen but this can be mitigated against with several relatively small changes.

Many applications will be largely unaffected, amplicons in heterogenous samples are likely to suffer the most.
Send your samples with different barcodes each time if you can.
Make sure to ask your facility to perform the washes between runs.

There is a hack to fool the MiSeq into thinking it is clean, don't use it! A maintensance wash is a bit of a pain but we'll be running them until this is fixed.

Okay, that makes sense as to the source of the inter-run cross-contamination. Does the hi-seq operate in the same fashion?

Still, that seems like a bit of a design flaw. Why would an engineer design a non-disposable piece of tube for the loading on the chip? For that reason alone I would be hesitent to use this in the clinic.

-Tom

[mcnelson.phd]

Man, I need to check these forums more regularly. We've been seeing this problem in all of our amplicon runs. At first we assumed it was contamination during the sample prep, but after running two completely distinct amplicon types (16S and non-ribosomal gene amplicons made using very different fusion primers) we still found low level mixing of the two despite the samples never coming anywhere near each other except that they were both run on the same MiSeq.

I know my FAS said that the MiSeq tubing can't stand repeated exposures to KOH, so I guess that's out of the question as a cleaning solution, but hopefully they'll come up with something.

[thomasblomquist]

Just how non-disposable is this tubing? I'd be willing to pay an extra $100 a run just to have contamination free assurance.

Can this topic be forwarded to someone at Illumina that is responsive to this need?

-Tom

[mcnelson.phd]

Just how non-disposable is this tubing? I'd be willing to pay an extra $100 a run just to have contamination free assurance.

Can this topic be forwarded to someone at Illumina that is responsive to this need?

-Tom

I'd say that the tubing is very non-disposable because from what my FAS and one of Illumina's field technicians have told me, in order to replumb the MiSeq you're looking at taking the whole system apart, including the optics, and would then have to put it all back together and re-align the optical table. My FAS said he once nicked one of the lines while doing an optical adjustment, and it took him an entire day to get just the one line replaced (I don't know which it was) and put the optics back in alignment.

Basically, Illumina will need to put out a new recommendation for a solvent that won't destroy the lines but will also sufficiently clean them to fix this issue.

[thomasblomquist]

And the hiseq?

-Tom

[james hadfield]

The tubing inside the MiSeq goes from the cartridge via a fixed sipper, through tubing to a vici valve, and through a final tube to the flowcell. It does not appear to be easily replaceable.

However a better wash is likely to decrease the carry-over.
Washing more frequently may well help.
Changing barcodes between runs should help.

These are all things we are in control of.

I think HiSeq 2500 comes with the option of cBot clustering if this is the case then there is almost no issue for 2500.

[kameran]

Hi everyone, this is Kameran and I’m part of the Illumina Technical Support team. Illumina has posted a bulletin, “Best Practices for High Sensitivity Applications: Minimizing Sample Carryover”, which can be found by following this link:
https://icom.illumina.com/MyIllumina...applications-m


In summary, run to run sample carryover is more likely to have an effect on very low detection threshold applications. Our internal testing found that, when the instrument is properly washed and maintained, sample carryover typically remained below 0.1%, representing 1 read in 1000. Carryover may arise from a number of steps in the sequencing workflow, including incomplete removal of template DNA before the next run and various steps of library preparation. Proper maintenance procedures with water and Tween are recommended to minimize carryover. We currently do not advise the use of other wash additives, such as bleach, Triton or other decontamination solutions.

Of course, we’re always available by phone or email ([email protected]) to discuss any additional questions or concerns.

[thomasblomquist]

Hi everyone, this is Kameran and I’m part of the Illumina Technical Support team. Illumina has posted a bulletin, “Best Practices for High Sensitivity Applications: Minimizing Sample Carryover”, which can be found by following this link:
https://icom.illumina.com/MyIllumina...applications-m


In summary, run to run sample carryover is more likely to have an effect on very low detection threshold applications. Our internal testing found that, when the instrument is properly washed and maintained, sample carryover typically remained below 0.1%, representing 1 read in 1000. Carryover may arise from a number of steps in the sequencing workflow, including incomplete removal of template DNA before the next run and various steps of library preparation. Proper maintenance procedures with water and Tween are recommended to minimize carryover. We currently do not advise the use of other wash additives, such as bleach, Triton or other decontamination solutions.

Of course, we’re always available by phone or email ([email protected]) to discuss any additional questions or concerns.

Thank you Kameran. What is the rate of decrease over multiple uses and washes? Specifically, using a shared instrument, how many other samples should I interdigitate between returning to previously used barcode combinations to achieve a 1:10,000; 1: 100,000, etc. carry-over?

-Tom

[kameran]

Hi Tom,



Studies regarding barcode integrations and degradation rate for run to run carryover have not been performed. However, customers have reported steady decreases from wash to wash. Anecdotal evidence suggests that, if carryover was 1 in 1000, it would be more like 1 in 1000000 the next time. This assumes there isn’t template that has dried and gotten “stuck” somewhere in the system. The more rinsing we can do on the MiSeq, the lower the carryover rate is expected to be.

[Harlon]

Hi Tom,

Studies regarding barcode integrations and degradation rate for run to run carryover have not been performed. However, customers have reported steady decreases from wash to wash. Anecdotal evidence suggests that, if carryover was 1 in 1000, it would be more like 1 in 1000000 the next time. Thlis assumes there isn’t template that has dried and gotten “stuck” somewhere in the system. The more rinsing we can do on the MiSeq, the lower the carryover rate is expected to be.

Hi Kameron

Thanks for the link to the technical bulletin. We had already implemented all of the fixes you suggest, including a maintenance wash between every run, prior to reporting our carry-over problem.

We have noticed that carry-over disappears within 2-3 runs at the read depths we are doing. Suggesting that additional washes would help. That said, I am really hoping Illumina will come up with a better solution than 'run 2 maintenance washes between runs'. Aside from the inconvenience, it increases the total time for a 150bp paired end run + wash to >24 hours. Since we're running our MiSeq pretty much every day, this would be a big hit to productivity for us.

Cheers
h

[thomasblomquist]

This message string has been very helpful for our specific purposes. So, I'm thinking that if we have three batches of barcoding reagents that we rotate through that should be sufficient amount of inter-run washes to eliminate the tubing contamination of barcoded product from a previous runs that used the same barcode. I.e. every third to fourth run may have the same barcode.

Thanks again. Very helpful.

-Tom

[nhunkapiller]

Is there a consensus as to whether this issues affects the HiSeq 2500 in Rapid Mode?

[james hadfield]

The answer is in Kameran's post, although why Illumina only put this on iCom and not out there publicly is a little galling.

Hi everyone, this is Kameran and I’m part of the Illumina Technical Support team. Illumina has posted a bulletin, “Best Practices for High Sensitivity Applications: Minimizing Sample Carryover”, which can be found by following this link:
https://icom.illumina.com/MyIllumina...applications-m


In summary, run to run sample carryover is more likely to have an effect on very low detection threshold applications. Our internal testing found that, when the instrument is properly washed and maintained, sample carryover typically remained below 0.1%, representing 1 read in 1000. Carryover may arise from a number of steps in the sequencing workflow, including incomplete removal of template DNA before the next run and various steps of library preparation. Proper maintenance procedures with water and Tween are recommended to minimize carryover. We currently do not advise the use of other wash additives, such as bleach, Triton or other decontamination solutions.

Of course, we’re always available by phone or email ([email protected]) to discuss any additional questions or concerns.

They say "To avoid direct introduction of samples on the HiSeq 1500/2500, the TruSeq® Rapid Duo cBot Sample Loading Kit (CT-402-4001) can be used to load samples onto the flow cell. The Rapid Duo kit contains disposable parts, thereby limiting the opportunity for carryover during clustering."

I think it is clear there is a fail-safe work-around for HiSeq 2500. Now we just need something a little better for MiSeq and we're all happy once more!

[id0]

I am just trying to understand this issue more. If the contamination is less than 1%, it seems largely irrelevant. Based on my experience, trying to call variants of that frequency will yield a huge amount of false positives. Isn't the sequencing error rate a much bigger problem that makes this a moot point?