Why would I buy an Ion Torrent?

[Haiqu]

I've been looking at used Ion Torrent sequencers on eBay lovingly, then researching to ensure that I don't buy yet another gadget I can't use. So here's what I've discovered so far:

1. They need a dedicated server and several extra bits like the One Touch, so they're useless alone. The full setup is about $10k used.
2. Cost of use includes $90 for the cheapest chip, plus $300 in chemicals. That ain't cheap when you consider that I bought a WGS test on special from Dante Labs for $200
3. There's no information publicly available from Thermo Fisher Scientific. In fact when I visit their website, I'm blocked by a message that tells me these products aren't available in Australia. That, to me, seems stupid in the extreme.
4. The chips can be cleaned and reused indefinitely if you're willing to put up with minor contamination (83% clean is what I've been told.)
5. They are only really useful for small jobs. Doing a whole human genome would take forever and use about 1,000 chips.
6. Reagents have a shelf life of 3 months.
7. Sample preparation is time consuming and complex.

So the above indicates major suckage for the hobbyist.

Can anyone suggest a better alternative? I'm happy enough with a system that can only sequence a small genome but RNA profiling would also be nice too. Where is Moore's Law taking us? And is there something just around the corner that is both affordable and available for the low end of the market?

[r.rosati]

4. The chips can be cleaned and reused indefinitely if you're willing to put up with minor contamination (83% clean is what I've been told.)

Hi. Do you actually have a protocol for that? I've tried, unsuccessfully.

[Haiqu]

You'd have to ask this guy: https://hackaday.io/project/160183-d...uencer/details

[cmbetts]

Is there a reason you need your own sequencer? You're highly unlikely to ever break even vs making homebrew Illumina libraries and sending them to a service facility that will pool them onto a HiSeq/NovaSeq.

My old company hated running our PGM. It's very user unfriendly.

If you're just interested in dinking around at home, go with Oxford Nanopore.

[Haiqu]

I live in Australia. Such services don't exist here.

Oxford Nanopore was considered and rejected on two bases: 1. Cost and short life of reagents. 2. The proprietary nature of the system. Sensors need to be returned to the manufacturer after 12 weeks. i.e. You don't actually own what you pay for.

So far no-one has come up with a system that's usable for the hobbyist. Everything seems to be aimed at batch processing. It's like computing in the 70's.

BTW I don't see this as a problem, I see it as a business opportunity.

[GenoMax]

@Haiqu: You should clarify if you are looking at Ion as a hobbyist (outside of a research institution) i.e. for DIY use.

[Haiqu]

My intro in the Introductions section clearly states that I'm retired and consider this a hobby activity. Sorry if I seemed too knowledgeable for that to come through here, it's a side-effect of being smart.

But I would have thought that my being interested in eBay equipment that's almost 10 years old would have given the game away.

[Bukowski]

Originally Posted by Haiqu

My intro in the Introductions section clearly states that I'm retired and consider this a hobby activity. Sorry if I seemed too knowledgeable for that to come through here, it's a side-effect of being smart.

No one digs through your post history to answer your questions.

The Ion Torrent workflow is notoriously finicky. Many people submit DNA or pre-made libraries to service providers in other countries in order to get their sequencing done. And outside of professional contexts too. The economic feasibility of this has already been outlined, let alone the issue of tying yourself to a dead platform (I'm interested in how you would source reagents to run a PGM).

Are you sure a MinION and some Flongles wouldn't be a more cost effective route? I bet there's a lot of DIYbio people floating around that would ameliorate the cost of a starter pack. And the library prep workflow is a lot less complex.

[nucacidhunter]

Originally Posted by Haiqu

I live in Australia. Such services don't exist here.

There are several companies in OZ that will sequence your own prepared libraries. You can PM me if you one a list.

[Haiqu]

Originally Posted by Bukowski

No one digs through your post history to answer your questions.

Sheesh, another one. From the top post in this thread:

"So the above indicates major suckage for the hobbyist."

The Ion Torrent workflow is notoriously finicky. Many people submit DNA or pre-made libraries to service providers in other countries in order to get their sequencing done. And outside of professional contexts too. The economic feasibility of this has already been outlined, let alone the issue of tying yourself to a dead platform (I'm interested in how you would source reagents to run a PGM).

All the ancillary stuff is available via eBay. But that's a Bad Idea (TM) for the same reason that outsourcing to overseas companies is a Bad Idea. Lead times are going to stretch into infinity. I'm still waiting (since February) for a WGS I did with an overseas company. Not acceptable.

Are you sure a MinION and some Flongles wouldn't be a more cost effective route? I bet there's a lot of DIYbio people floating around that would ameliorate the cost of a starter pack. And the library prep workflow is a lot less complex.

It seems to me neither option is a cost-effective route. Added to that, the MinION is very proprietary, sensors need to be returned to the manufacturer.

[Haiqu]

Originally Posted by nucacidhunter

There are several companies in OZ that will sequence your own prepared libraries. You can PM me if you one a list.

Awesome!! Will PM you, thanks. They sure don't advertise.

Edit: Oops. Forum doesn't seem to have a PM system. I've emailed you, hope it gets through.

[nucacidhunter]

Following is the link to some OZ sequencing service providers:

http://www.agrf.org.au/contact-us

https://www.ramaciotti.unsw.edu.au/

http://dna.med.monash.edu.au/

I am sure they all will do good job.

[Haiqu]

Thanks nucacidhunter. I did also receive the email.

I've also become aware of the unreleased ONT SmidgION device. If they can get the cost of use down it may prove to be a viable alternative, although they (stupidly) haven't announced software for a PC and seem to be aiming at tethering the thing to an iPhone. They also screwed up by using the wrong gender of mini-USB connector.

The ONT technologies are of most interest so far, although I believe the costs need to come down by another order of magnitude before broad use ensues from hobbyists. They need to decouple reagents from the package so that individuals can source them locally, bring down the cost of the flow cells and forget about their enforced use of commercial data analysis services like Epi2Me. Only then will I believe their hyperbolic claim of "Anyone, anywhere, anytime" usage.

[gringer]

The MinION flow cell return to ONT is so that they can recycle them for use in training and internal research. If you contaminate the flow cells such that they are not safe for posting back, you don't need to return them. There is no return required for flongle flow cells.

There's no need to use Epi2Me; the only necessary proprietary software is the MinKNOW sequencing software (for controlling the MinION), which produces HDF5 files containing raw signal traces from the sensor arrays. You're free to design your own base-calling software (e.g. Chiron) and bypass every other part of the process. As far as I'm aware, all other sequencing companies also have proprietary software for controlling initial data acquisition.

Minimum per-sample cost via 12-sample multiplexing with the rapid barcoding kit and a flongle flow cell is about $9 USD. The only thing cheaper than that which I'm aware of (possibly excluding microarray) is a single Sanger sequence by service, or do-it-yourself Sanger.

[Haiqu]

If the cells can be recycled, that option should be available to the purchaser. It has already been discovered that cells from the older Ion Torrent machine can be used indefinitely, and I suspect ONT don't want the user experimenting with that possibility. Their official reason per their website varies from "Please return your flow cells after use." (i.e. no reason) to "It is a requirement of the contract." depending on where you read. In any case, I can't see any valid reason to be paying return shipping to the UK if the device is merely going to be disposed of.

Good to know there are options for base calling. My objection to Epi2Me was the necessity for large online data flow, which is difficult when you're mobile and expensive in general in Australia. As an example, I had to reach out to a friend who works for an ISP to merely download a 100GB BAM file, because such bandwidth isn't available to the end user at any price as yet. Being a kiwi, I'm sure you can relate.

I can't see a hobbyist needing to handle the volumes for 12-sample multiplexing, and the learning curve for barcoding would complicate an already complex procedure. A more likely scenario is use and rewash of the flow cell, where maybe 4 or 5 samples can be read at best. With single flow cells being $900 (plus reagents at several hundred bucks) the cost will be more like $300 per sample, especially when you take into account that they have a limited (12 week) shelf life.

All this just outlines the differences between commercial operation and home use, and is no criticism of ONT technology per se. There is no product for the low end of the market at present, and the one that might have fit the purpose - the SmidgION - has been stillborn, judging from the lack of progress since 2016.

I fully expect something new to eclipse ONT soon anyhow, given that it is now fairly old technology in "dog years."

[gringer]

Originally Posted by Haiqu

In any case, I can't see any valid reason to be paying return shipping to the UK if the device is merely going to be disposed of.

Customers don't pay return shipping to the UK. That cost is met by ONT, even for flow cells that are being returned from New Zealand. I suppose if you want to be cynical, you could say the return shipping cost is included in the price of the flow cells.

I can't see a hobbyist needing to handle the volumes for 12-sample multiplexing, and the learning curve for barcoding would complicate an already complex procedure.

You can either multiplex and reduce the per-sample cost, or not multiplex and get a higher yield per sample. Flongle flow cells ($100 USD each, for about 1/10 of the yield) make that decision a bit more palatable. There's essentially no learning curve for rapid barcoding on top of standard rapid sample prep. Here's the normal sample prep:

1. Add 2.5μl fragmentation mix to 7.5μl sample in a sample tube
2. Heat to 30°C for 1 min, then 80°C for 1 min
3. Add 1μl adapter mix, Wait 5 mins
4. Mix with sequencing buffer and load onto flow cell

Here's the barcoded sample prep:

1. Add 2.5μl barcoded fragmentation mix to 7.5μl of each sample in a sample tube
2. Heat to 30°C for 1 min, then 80°C for 1 min
3. Mix all the prepared samples together
4. Put 10μl into a new tube
5. Add 1μl adapter mix, Wait 5 mins
6. Mix with sequencing buffer and load onto flow cell

Sure, it's two more steps, but it doesn't involve any more learning.

A more likely scenario is use and rewash of the flow cell, where maybe 4 or 5 samples can be read at best.

Yes, that can be done as well. ONT provide a wash kit for $190 that works for 12 rewashes, but people have discovered (with ONT's help) that a nuclease flush is actually more effective (carryover with a nuclease flush is something like 0.01% vs 1-5% with the wash kit).

It works. Depending on how much yield you want, a flow cell can be washed and re-used for 8 runs, probably more.

However, the flow cells contain ions that pass from one side of the membrane to the other, and these deplete over time through using the flow cell. Due to this, sequencing is more effective (i.e. produces a higher yield) when a flow cell is fresh, so you will get a much higher yield from 8 samples multiplexed, than from 8 samples washed and run one after the other. There's also a bit of a risk when washing a flow cell of damaging the sequencing membrane.

[that's one of ONT's protected technologies - a polymer membrane that is transport-stable, with nice electrical properties that make it easy to load single nanopores and hard to load more than one.]

With single flow cells being $900 (plus reagents at several hundred bucks) the cost will be more like $300 per sample, especially when you take into account that they have a limited (12 week) shelf life.

It depends on the application. A flongle flow cell with a rapid barcoding kit requiring low yield (50-100 Mb / sample) works out to about $9 / sample [as I've already mentioned]. A single whole genome sample (5-15 Gb) with a MinION flow cell and 1/6th of a ligation kit is $1000 USD for ONT reagents, plus another $100ish for NEB reagents and Ampure beads.

The 12-week shelf life is ONT's warranty period. Many ONT community users (including me) have found that flow cells will last for over a year. I have done washes of used 6-month-old flow cells with well-expired rapid kits, and got enough sequence out of the flow cells to assemble viral genomes.

For some people, I appreciate that the cost of a $100 flow cell will be beyond them, and they may not be able to take 1/144th of a rapid barcoding kit from someone who has some spare. But saying that's completely out of the price range of a hobby / home user seems a bit odd to me.

[Haiqu]

Originally Posted by gringer

Customers don't pay return shipping to the UK. That cost is met by ONT, even for flow cells that are being returned from New Zealand. I suppose if you want to be cynical, you could say the return shipping cost is included in the price of the flow cells.

The return shipping price is built into that exorbitant $900.00 per cell.

You can either multiplex and reduce the per-sample cost, or not multiplex and get a higher yield per sample. Flongle flow cells ($100 USD each, for about 1/10 of the yield) make that decision a bit more palatable. There's essentially no learning curve for rapid barcoding on top of standard rapid sample prep. Here's the normal sample prep:

1. Add 2.5μl fragmentation mix to 7.5μl sample in a sample tube
2. Heat to 30°C for 1 min, then 80°C for 1 min
3. Add 1μl adapter mix, Wait 5 mins
4. Mix with sequencing buffer and load onto flow cell

Here's the barcoded sample prep:

1. Add 2.5μl barcoded fragmentation mix to 7.5μl of each sample in a sample tube
2. Heat to 30°C for 1 min, then 80°C for 1 min
3. Mix all the prepared samples together
4. Put 10μl into a new tube
5. Add 1μl adapter mix, Wait 5 mins
6. Mix with sequencing buffer and load onto flow cell

Sure, it's two more steps, but it doesn't involve any more learning.

Doesn't sound too bad, but you're glossing over the process needed to barcode the samples. That might well be simple too, but from here I can't tell. It's also a trade-off between number of samples and test resolution. I can't see 800 cells / 12 barcoded samples being particularly useful except perhaps for identifying well-known small virii. I expect to be doing de novo plant work at first, and from what I've seen of other people's work online, they go through a full MinION cell per test and still have trouble piecing together a complete sequence.

But as I've said, I am a noob, albeit a determined one. I'll get back to you in three months when I've completed the MIT 7.00x Introduction to Biology course I just started. After which I'll be doing the 7.QBWx Quantitative Biology Workshop, to be able to handle the data crunching we already discussed.

The 12-week shelf life is ONT's warranty period. Many ONT community users (including me) have found that flow cells will last for over a year. I have done washes of used 6-month-old flow cells with well-expired rapid kits, and got enough sequence out of the flow cells to assemble viral genomes.

From what I read, they have to returned if not used within 12 weeks, but I can't imagine anyone taking that seriously. They certainly need to be tested within that timeframe for warranty coverage.

For some people, I appreciate that the cost of a $100 flow cell will be beyond them, and they may not be able to take 1/144th of a rapid barcoding kit from someone who has some spare. But saying that's completely out of the price range of a hobby / home user seems a bit odd to me.

The US$90 flow cell is used with the Flongle, which is another US$1,860 on top of the cost of the MinION. So now, with chemicals, we're up well over three grand for that first sample test. That's why I'm so disappointed that the SmidgION isn't available yet, it's an all-in-one unit. However I don't expect it to be very useful, and the value for money will always with the unit that handles the most workflow.

[Haiqu]

Just following up on the SmidgION ... it will not be available for at least two years, according to Oxford Nanopore.

[Haiqu]

Originally Posted by r.rosati

4. The chips can be cleaned and reused indefinitely if you're willing to put up with minor contamination (83% clean is what I've been told.)

Hi. Do you actually have a protocol for that? I've tried, unsuccessfully.

Did you ever get a response from that Russian guy who built his own sequencer with the Ion Torrent chip? I tried and have had no reply after three weeks.

[scrosby]

Originally Posted by Haiqu

Did you ever get a response from that Russian guy who built his own sequencer with the Ion Torrent chip? I tried and have had no reply after three weeks.

I just sent him a note about this which is pretty cool!
https://twitter.com/bryancmounce/sta...01646840258560