Buy Ion Proton for core facility or not?

[flxlex]

Hi all,

I represent a sequencing cor facility (Oslo, Norway) of moderate size with:
- one Illumina HiSeq
- one MiSeq
- two Roche/454 GS FLX, one upgraded to GS FLX+
- one Ion Torrent PGM
- one PacBio RS

We have been additionally granted an amount which allows us to buy an Ion Proton, but are really uncertain if this is a smart move for us at this point:
- the HiSeq is running 24/7, we may be getting a second one
- the MiSeq is a problematic machine to get to work stably, but will probably generate some user interest
- the Ion PGM is collecting dust
- the 454's are mainly used for amplicon sequencing, the FLX+ finally seems to work and may generate some interset
- the PacBio will take some convincing potential users, but we have some 'local' research projects where we use it for the extra long reads, which turn out to be very helpful (de novo genome assembly of complex genomes)

So, how would the Ion Proton fit in:
- easy labwise, the crew is really fed up with emPCR
- once the 400 bp reads are there, useful for amplicon sequencing
- fast turn around time for intermediate sized projects
- perhaps it could be used to replace the 454 long insert mate pairs (longer than the 10kb possible on the Illumina); so far, there only is a demonstrated protocol, though.

Why would it be a bad idea to buy one:
- maybe it is a sign on the wall that the PGM is collecting dust
- you can get the same, if not cheaper and with better quality, done on the MiSeq/HiSeq2500_rapid_run_mode
- so, a big risk we will not generate a user base for it

We are really uncertain here. Any opinions? Would you buy a Proton today if you had the funds?

[NextGenSeq]

Our MiSeq works flawlessly even with the new upgrade

[GenoMax]

Looks like you have one of everything .. which is not a bad thing. Getting techs to learn intricacies of each platform is challenging. Perhaps you have enough man power available to dedicate to diverse instruments.

I am not sure what kind of problems you are having with the MiSeq (aside from the outstanding low nucleotide diversity affecting quality issue ours has been a stable workhorse).

Real question to ask is are there projects on the horizon that would suite a Proton better than existing infrastructure. If the answer is no then investing in a second HiSeq (with the 2500 upgrade) may be more practical.

[scbaker]

I'm a bit confused by this statement:
"easy labwise, the crew is really fed up with emPCR"

Are you saying that the Proton doesn't require emPCR? My understanding is that the PI and PII chips WILL require emPCR. Avalanche, which will eliminate emPCR, won't be available for another 16+ months and will only work on PIII chips.

[NextGenSeq]

I hate emPCR too. I was told that even when they eliminate emPCR they will still be doing the chemistry on beads. Not sure what that means. Maybe a chemical bridge amplification on beads instead of a flowcell?

[MrGuy]

I think it is strange people keep comparing 454 emPCR experience with Ion Torrent emPCR experience. I used both and utterly hate the 454 emPCR and don't wish that on anyone. The people who took the 454 from us don't like it much either.

The PGM, nowadays, is "not so bad" to use. Mind you, you still need to use the one touch and an ES for the time being. The v1 onetouch is ok, but not great. It's been greatly improved since day one, but am looking forward to v2 or chef. Illumina workflow is simpler as it stands today, but too black box for my liking. Chef and later Avalanche should change that. Chef should work on your PGM and your Proton, if you end up getting that.

I think the biggest improvements over the past year were:
- the DL upgrade on the one touch
- 300 bp chemistry has a good error profile, but I don't really need those long reads except for amplicon. I tune it back a few hundred to speed the runs.
- nitrogen instead of argon has greatly improved all the pH/chip issues people used to see. Nitrogen is a better controlled gas than argon from a manufacturing perspective. Also cheaper and easy to plumb.
- the fact that you *can* run two chip runs in a working day. I've done it with two 300 bp runs. Mind you the longer chemistries take longer to run, obviously, but the first will finish before you can start the next run and go home to the family.

Avalanche, from what I understand, will be working on all ion platforms. Whether or not it is bead based or wildfire or whatever, I don't really care. I'm more interested in the fact it will be drastically faster than emPCR or bridge amplification (cBot/Illumina).

My suggestion, since you already have it, is to just try the PGM out. It'll cost you a few runs, be prepared for a few bumps, but if you have pro 454 guys, the PGM will be a walk in the park. You'll be blowing through all the specs from day one.

I would guess the experience will be similar from PGM to Proton, but keep in mind the Proton *is* a new chip and a new machine (tired pun intended). There will be some bumps along the way as you experienced with the MiSeq which was heralded as the mini HiSeq and just splendid. Companies are generally interested in making a sale. You are interested in doing good science in the most cost effective manner. I'm guessing you can use this to your advantage, especially considering your big fish you guys just landed.

Regardless, I am of the opinion that NGS platforms in and of themselves are extremely boring sequence readers that have oddly become a fashion statement. Library prep and interpreted analysis is interesting as it is there you get to ask your questions and get your answer. As long as the sequencer gives you the data you need for the price and in the time you need, then go for it.

[epistatic]

I have strong interest in a Proton but would not buy one yet.
A) Site in my state bought one, it arrived and is sitting in a box until it can be installed in December. What are install specs?
B) I would like to see data produced on the Proton, not just told that it is the PGM with more yield.
C) What is the Q-score spec for the reads being produced? What is the amount of data after quality filtering, read trimming, etc?

[scbaker]

I'm a bit confused by this statement:
"easy labwise, the crew is really fed up with emPCR"

Are you saying that the Proton doesn't require emPCR? My understanding is that the PI and PII chips WILL require emPCR. Avalanche, which will eliminate emPCR, won't be available for another 16+ months and will only work on PIII chips.

Nope, I got that wrong. Life says Avalanche will work with "all Ion systems" (via In Sequence).

[Jeremy]

If you already have a HiSeq, a MiSeq and a PGM then a Proton with the PI chip will probably not get used. But, when the PII and PIII chips come out you might notice some customers using the Proton instead of the HiSeq.

The reported results from the Lifetech conference was that paired end on the PGM (and I guess the Proton also) basically doesn't work, which IMO negates its benefit of longer reads compared to HiSeq.

Unless you need to spend the money right now, I would wait and see how the proton functions when the PII chip is released and the reads get longer. At the moment the Proton is a lot of promises and very little data to back it up.

[flxlex]

Hi all,

Thanks for the answers, really helpful. Although no consensus yet :-)

A few replies:

- yes, with emPCR I meant the 454 one.
- our Miseq, and replacement MiSeq are not performing well. I've heard rumours that we're not the only ones, although an are very happy with it.
- unfortunately, we need to decide what to use the funds on by year's end
- there is indeed no Proton data available through the Ion community, strangely enough... They really need to get some out there.

Keep your comments coming!

[gojohn10]

If not proton, then what? What are the the other options for these funds? A different NGS sequencer? Something else entirely? If the question is Proton or nothing, then the choice is obvious. Otherwise, it might be best to wait a bit so you can get a better idea of whether a Proton will fit a niche in your lab. I suggest waiting only because it sounds like you aren't in desperate need of another sequencer at this moment.

[flashton]

Hi Lex,

Did you get a Proton in the end? If so, why? If not, why not?

[Buzz0r]

it might be worth waiting till they actually can produce some data..meanwhile the news from Illumina about new MISeq development in 2013 and the new Moleculo acquisition for long reads might be a sweetener since you already have MiSeq ready on site. Are they performing better now or do you still have problems?

[flxlex]

We did in fact ended up buying a Proton, believe it or not. There were many factors, some strategic, that made us decide this. Time will tell whether the addition was a smart one...

The MiSeq has been replaced and the new one seems to be doing fine.

[GenoMax]

We did in fact ended up buying a Proton, believe it or not. There were many factors, some strategic, that made us decide this. Time will tell whether the addition was a smart one...

Have you started using the proton? Initial impressions?

The MiSeq has been replaced and the new one seems to be doing fine.

Once it is working .. MiSeq is truly a well designed self-contained workhorse that is well engineered.