Ion Torrent and desktop sequencing: The Death of the Core

[genbio64]

Greetings everyone,
With all of the new exciting announcements involving the announcement of $1,000 genomes I assume people will scrutinize the NGS field closely. With both Illumina and Ion Torrent taking the early leads it seems the lines in the sand have been drawn. But Illumina's genomes must be done on the massive HiSeq platform while Ion Torrent is reducing machine size and price. With the financial barrier of entry being so dramatically reduced, and under the assumption it will continue to fall, what will be the future role of the Core facilities if everyone can afford their own sequencers?
Granted this has occurred with other technology (i.e. PCR, oligos, etc) this field has been so disruptive that it does beg to ask the question.

[chadn737]

I'm sure those labs that do nothing but sequencing will purchase their own machines, but I am willing to bet that the majority of labs only want to do the occasional sequencing project (such as ours). Such labs may only do next gen sequencing project once a year at the most. Which does not justify paying 10s of thousands or hundreds of thousands to buy a machine.

[GW_OK]

There's also the per-sample cost to consider. A Hiseq can do more for less than an ion torrent or a miseq.
I think chad's point is the most salient, though. Perhaps we'll see a proliferation of "cores", but with miseqs and ion torrents instead of the big guys.

[GenoMax]

The question ultimately is of best practice. As indicated by Chad it would not make sense for casual users to invest time/energy in picking up the expertise needed for running these type of analyses. Until the cost of these machines drops below $10K they would still be "capital" investments. In this case, using them infrequently would not be the best use of research money.

Analyzing the data should be the focus of researchers rather than generation/management of the data. That part is best left to the pros in a "core".

[DNADEB]

From the standpoint of 16 yrs as a Core Director at a university, the cost of populating the average lab with the equipment to do assess and prepare the samples puts in only in the realm of a few high level labs. That isn't even addressing the problems with a swarm of untrained undergrads, grads and even techs trying to do the work. And on top of that, these machines break down all of the time and need constant attention. Possibly with a dedicated tech or lab manager per machine, the work could get done but not like a core. Also, the cost of unused reagents will become a factor.

[epistatic]

I see many investigators having a hard enough time providing suitable DNA and RNA for sample prep. Very few labs have the desire to make sequencing libraries even if they have the ability. Any instrument with emulsion PCR also adds a huge layer of complexity to the process and needs well trained staff. If one of the nanopore instruments can work without sample prep that would be an easier entry point but analysis still would be the main issue. Most don't want to make libraries, can't analyze their own data and see no point in being the middle man with a benchtop sequencer. It is easier and more economical to pay for sequencing data than purchasing and operating capital equipment. The reps always forget to talk about service contracts and ancillary equipment needed with these benchtop boxes.

I think generating the data will be cheaper, but then the investigators need that data to be in a biologically interpretable format. Core facilities will need to gear up more for analysis than the current focus on data generation. Price per sample isn't going to get cheaper when you include the analysis and it is hard to provide analysis without a charge-back. Many want to work on a collaborative model but that doesn't carry a core too far. The marketing spin is making people think sequencing is getting really cheap but it will sit at a stable not so cheap price-point for awhile when you factor in the all inclusive cost of getting the data.

[DanFrost]

From the perspective of our clients (core directors and customers of multiple cores), the role of the core is changing from "the people who do it all" to "the people who understand this the best."

As technology in both data generation and data analysis matures, defining best practices and workflows will be the role of the core. Very in depth, very technical, and requires maximum expertise. Running samples on smaller machines in a lab will become more common place, and data analysis will become more standardized and accessible from all levels of researchers.

Will it kill the core? Absolutely not. Will you top tier folks be happier not repeating everyday processes over and over and over again? Probably. Sounds like a nice future to me.

[ETHANol]

I don't see it happening. Much more likely that companies like Macrogen will take over like they have with Sanger sequencing or say IDT with oligo synthesis. These are both things cores use to do. Cores will move on to something new.

As for the data analysis, you can't efficiently source that out. There is too much disconnect between the people that understand the experiment and the people analyzing the data. Cores will continue to do this for biologists that are too lazy to learn how to analyze their own data (with an emphasis on lazy).

I don't really see much in the IonTorrent or MiSeq for your average lab. The per base or per read cost is much higher then the HiSeq. Why do it in your lab for more money then it would cost to have someone else do it for you?

[skycreative]

There is too much disconnect between the people that understand the experiment and the people analyzing the data. Cores will continue to do this for biologists that are too lazy to learn how to analyze their own data (with an emphasis on lazy).?

Agree with you