Ekaterina,

I think you are on the right rack with your troubleshooting. It sounds more like library prep than a bad chip to me.  Yes, a contaminant in your sample prep could cause the library prep to fail.  Do you run an Agilent assay before library prep to get a RIN number for your total RNA?  What kit do you use to make total RNA?  This can also cause problems. Finally, what was the A260/A230 ratio for the samples?  Low 260/230 can mean there are contaminants in your sample which can inhibit downstream library preparation.

Hope this helps!

LTR

P.S. I suggest you personally assess the quality of the samples after sequencing using FastQC on the GALAXY suite.  Very easy to use and you will understand why so many reads were cut out of downstream analysis.

Thank you, Lawrence,

I suppose that total RNA was extracted by TRIZOL reagent. Before constructing the libraries I estimated the quality of total RNA using Agilent RNA Nano Chip, and I've seen 18S and 28S peaks clearly. (I couldn't calculate RIN number, though, because the Agilent gel was expired and the peaks weren't on their expected migration times and weren't calculated automatically, and the "manual integration" option in the Agilent software was for some reason inactive).

I used NEBnext kit for rRNA depletion and then Ion Total RNA-Seq Kit v2 for library preparation. A230 in the final libraries was very high, much higher that A260. But I've seen such spectrophotometry results before and the sequencings were good then. I also asked my colleagues about this and I was told that such impurity doesn't affect sequencing results. I suppose that high 230 nm absorbtion in my libraries has something to do with some traces of Magnetic Bead Purification Module reagents.

I looked on the test fragments on the chip - 50AQ17=85%, and the first peak on read lenght histogram is near 100 bp - I understand that the chip is good??

I have seen similar issues with the Ion Proton, and they were caused by a hardware issue with the machine. In our case, the front waste arm was causing a short inside the case, which was resulting in electronic noise leading to garbage reads. We had the arm retrofitted with a new one and the issues immediately went away.

I would recommend a few things. First, run a full chlorite clean on the instrument. It's possible there's a clog somewhere in the internal fluidics that is contributing. Second, run through the provided Ion Torrent control beads, which have a known sequence. If these come out fine, then the issue is further upstream, with your sample prep.

Next, I would look at the logs from your enrichment station. Are you using a stand-alone emulsion PCR machine, or the Ion Chef? With the stand-alone unit, it's important to pop off the case and check the levels of coolant - this can run low without the machine making a peep, and can ruin your PCR.

Also, did you check the beads that came off the enrichment station for % enriched? It's possible that process went awry somehow, so you ran a lot of beads without DNA fragments.

Jennifer, thanks,

I, too, think that the problem occured during emulsion PCR or enrichment.

We don't have Ion Chef and use One Touch 2 machine for emulsion PCR and One Touch ES machine for spheres enrichment. I have big qualms about the ES machine - while the PCR room is clean and has UV lamp, ES machine doesn't have any hood or lid, and while pipetting the sample stays in the open plate for half an hour, and last time I've clearly seen a tiny lint in the tube after enrichment. Maybe this lint contained DNAses that ruined my results. I guess I will wash ES with bleach and move it from the benchtop inside the PCR hood.

Regretfully, I don't know how to disassemble One Touch 2, and I even didn't understand what the "front waste arm" on Ion Proton is. Waste liquid flows down through the plastic nozzle to the plastic container, and I suppose that plastic parts can't  cause short circuit. Chip compartment of our sequencer looks almost exactly like in the picture attached and doesn't have a big metal lever like in PGM sequencer. Could you please give me any links to manuals explaining how to disengage these instruments properly?

We have Qubit fluorometer and we just ordered Ion Sphere Quality Control Kit, but we'll have to wait for its shipping at least for a month. But Torrent Server report says that 95% of wells with ISPs were live (with template), so I guess that empty spheres separation was done properly.

Since you see 95% of wells with live ISPs, it sounds like the OneTouch ES worked correctly. (We kept our ES stations on the open bench and never had problems.) I'd focus on emulsion PCR failure or sequencing failure. 

Check the saved logs on your One Touch 2 to determine if anything went wrong during the PCR. You should be able to save these onto a USB flash drive plugged into the back of the machine. This procedure is listed on p. 26 of the manual:

https://tools.thermofisher.com/content/sfs/manuals/MAN0014388_IonOneTouch2Sys_UG.pdf

The whole cover for the OT2's left side lifts off to give access the coolant tank. I've only seen it done by Ion Torrent field technicians and it doesn't seem to be covered in the user manual. I'd contact your FAS (field service agent) for this level of troubleshooting.

The "front waste arm" on the Proton is exactly what you think it is: it's where the liquid flows out of the machine into the plastic tank in front. This arm goes back further into the machine, and has metal internal components. That's what caused our issues. This isn't something you can investigate yourself. Questions about this should be directed to your FAS.

Thanks a lot!

I regard overhauling the instruments as a last resort, first I need to make sure that they indeed malfunction. I will clean all surfaces, change bottles and sippers on OT2 and other expendables that can be used several times, like cleaning chip, open up a new reagent kit. (As it is, I do chlorite cleaning of the sequencer every time before run, because we use it less often than every week, and I do MQ water cleaning after every run, so I guess that there can't be contamination in the internal fluidics).

I wish I can also download OT2 logs, but it says in the manual that logs are deleted every time when I turn off the instrument, and I did turn it off several days ago =( Also, the manual doesn't say exactly for what lines I should look specifically in the log - would it be simply "error" or whatnot, or would it be some numbers for which I need to know their optimal values.

If all this fails, I will call certified repairmen with a clear conscience.

Thank you one more time for your valuable suggestions!

Hi Ekaterina,

I suggest using the Ion Sphere Quality Control Kit because it helps in understanding if OT2 reaction failed or not.

I always measure our sample after OT2 step: the values should be between 10-20 % but I also loaded samples with lower (7%) and greater (30%) values. In the first case the chip loading was about 30% and the quality of reads was very low meaning that perhaps the PCR during OT2 did not work properly.

IIn the second case the chip was too full and I got a lot of polyclonality beads.

Erika, thanks for your help,

I didn't understand what parameter of sample do you estimate in %, could you please elaborate? Two ways of emulsion PCR quality control that I know about are Guava Flow Cytometer and Qubit fluorometer, but I never performed either of them myself.

Chip just can't be "too full": the maximal mathematically possible loading is 100%, but such loading is impossible to reach (and empty wells are used for calibration or something like this, so perfect 100% would be bad, too). Excess of polyclonal beads usually happens if you used too concentrated library in emulsion PCR: the optimal library concentration is 10-12 pM, with this concentration the majority of emulsion drops contain only one DNA template. However, even with optimal DNA concentration in emulsion there would always be 25-35% polyclonal beads because distribution of DNA molecules in drops is random.

Hi,

the optimal library concentration that we use for preparing OT2 reaction mix is 100 pM and not 10-12 pM. Which version of OT2 and sequencing kit are you using? We have the HiQ version but I am quite sure that also the previous v3 version required 100 pM as input for OT2 step.

At the end of OT2 you can measure the ISP sample using QuBit and Ion Sphere Quality Control Kit. You have to put the QuBit values in a spreadsheet provided by THermoFisher that calculates the "probable enrichment" of your sample. The enrichment value should be between 15-20 %.

If your sample has a lower % of enrichment maybe the OT2 reaction did not perform well.

Write me if you have doubts.

I use v3 kit. https://tools.thermofisher.com/content/sfs/manuals/MAN0009134_Ion_PI_Template_OT2_200_Kit_v3_QR.pdf (library dilution is on page 3)

Yes, it says 100 pM in the first line of the table, but next you add 9 volumes of water and so dilute it to 10 pM, and I meant this concentration (and then you dilute even further, of course, when you prepare the amplification mix). Earlier I used v2 kit and its manual explicitly says 10 pM in 100 ul volume - it's the same concentration and volume but described in different words.

We are still waiting for Ion Sphere Quality Control Kit. It only controls enrichment level? Can't you check unenriched spheres with it?

Hi Ekaterina,

the Ion Sphere quality control permits to check how many beads have ligated a DNA fragment. You perform this check before doing the enrichment phase with ES instruments. So the sample you measure is a unenriched one.

I do not know a possible check to control if also the enrichment process has worked well or not.

Have a look at this:

https://www.thermofisher.com/content/dam/LifeTech/migration/files/technical-reference-library/newsletters-journals/bioprobes/bioprobes-67.par.88529.file.dat/bioprobes-67-qubit-ion-torrent.pdf

An additional thing to check if the OT2 reaction has problems is to control the TF1 profile (Test Fragment) in the Torrent Server User Interface.

Thanks Erika,

So I understood that Qubit measures template adapter/ISP adapter ratio and helps you to tell if there was no or poor amplification and too little library-containing ISPs. But hypothetically, if amplification was successful but there occured any contamination or fragmentation of amplified DNA, I wouldn't be able to know this based on Qubit results.

I remember there's two steps in the protocol when you are recommended to check ISPs quality - before and after the enrichment step. But after the enrichment only Guava method is applicable.

I already checked TFs of this run, it said 50AQ17=85%, and the peak on read lenght histogram was near 100 bp. I'm not sure, but I decided that it means the positive control reads were good enough (based on this: https://biolektures.wordpress.com/2011/08/01/ion-torrent-test-fragment-accuracy-part-1/).

I must say to all the people who cared and answered here that I performed another emulsion PCR and sequencing after cleaning the room and replacing all the reagents by freshly opened ones, and the sequencing run was good.

So the problem wasn't caused by hardware, but I have yet to know which of the reagents gone bad.

Check this out, it's from the user guide.