I would contact your current Roche representative.  The press release of the end of life of the Roche 454 said that they will actually continue to make machines through 2015, and service machines through 2016.  There was no mention of how long they may continue producing reagents and kits though, or whether they will partner with anyone to make them available.  Ask your current area technical and sales reps about that, as they may know of plans to keep the reagent pipeline going (or not).  Presumably if they continue manufacturing through 2015, they must have some plans to keep those buyers in reagents for a reasonable service life of at least a few years (although likely nowhere near 5-10).

You certainly cannot stockpile reagents for 10 years as they simply do not last that long under the most ideal storage conditions.  But, if reagents will be available, then you could continue to use the machine.  The issue would be service if anything on the actual sequencer breaks, as parts will be ever more difficult to find, as well as people experienced enough to service the equipment.  Unlike other technology retirements (where companies retained staff to support both new and past generations), Roche is not replacing the 454 with anything (yet at least). It is laying off the employees and getting out of that particular business (although not out of the sequencing business entirely).

Honestly, however, expecting 10 years or more from a short lived technology that has now been discontinued and eclipsed by newer and better technologies is asking a lot.  The pace of progress and change in sequencing technology is just to fast too expect to get ten years total out of any current or recent technology.  Current mainstream technology is already set to be soon eclipsed by single cell sequencing technologies, which will mature rapidly in the next year or two.

So, I would say you should be fine at least through 2016, and for reagents maybe another year or two.  But I would use that time to also plan on retiring your 454 machine in the 2016-2017 time frame.

It is obviously not a problem to keep using this machine as long as you want as long as you have a stock of reagents. The sequence output will remain compatible with programs / algorithms. But, the real question I guess is why would you want to? Roche have a very good reason for dropping this platform!

I would contact your current Roche representative.  The press release of the end of life of the Roche 454 said that they will actually continue to make machines through 2015, and service machines through 2016.  There was no mention of how long they may continue producing reagents and kits though, or whether they will partner with anyone to make them available.  Ask your current area technical and sales reps about that, as they may know of plans to keep the reagent pipeline going (or not).  Presumably if they continue manufacturing through 2015, they must have some plans to keep those buyers in reagents for a reasonable service life of at least a few years (although likely nowhere near 5-10).

You certainly cannot stockpile reagents for 10 years as they simply do not last that long under the most ideal storage conditions.  But, if reagents will be available, then you could continue to use the machine.  The issue would be service if anything on the actual sequencer breaks, as parts will be ever more difficult to find, as well as people experienced enough to service the equipment.  Unlike other technology retirements (where companies retained staff to support both new and past generations), Roche is not replacing the 454 with anything (yet at least). It is laying off the employees and getting out of that particular business (although not out of the sequencing business entirely).

Honestly, however, expecting 10 years or more from a short lived technology that has now been discontinued and eclipsed by newer and better technologies is asking a lot.  The pace of progress and change in sequencing technology is just to fast too expect to get ten years total out of any current or recent technology.  Current mainstream technology is already set to be soon eclipsed by single cell sequencing technologies, which will mature rapidly in the next year or two.

So, I would say you should be fine at least through 2016, and for reagents maybe another year or two.  But I would use that time to also plan on retiring your 454 machine in the 2016-2017 time frame.

Thanks for your answers Michael and Russel. I guess I was pushing it a little when I said 10 years but in my lab we have 10 year-old PCR machines which do the job perfectly fine, why change them? I am guessing they will still work for few years. My reasonning is the same with 454. Roche stops it because it does not make enough money (which is fair enough). However, it works great for what I am doing. No other technology is delivering the same read length so there is no alternative at the moment. I actually don't have a 454 in my lab, but I am thinking of all these 'richer' labs who are/will get rid of theirs. I would be very happy to acquire a second hand one, providing that I can still use it beyond 2016.

> we have 10 year-old PCR machines which do the job perfectly fine, why change them?

A 10-year-old PCR machine performs the same function as a brand new one (for the same type of PCR), so there is not much reason to change it. A brand new PCR machine is not inherently different than an old PCR machine. Sequencing technology however is very different. It will take a very long time until technology matures to a point where a 10-year-old sequencer is as good as a new one. Also it's very unlikely that you would save money by buying and maintaining (!) an old 454. New, ever cheaper sequencers are on the horizon, ONP MinION is one example.

Also, I doubt anyone would take you seriously if you tried to publish some research findings in ten years from now which you made on a 454.

I can see your point Anton. As I said, I was pushing it a little when I talked about 10 years. But I think in terms of sequencing technology it is difficult to say that one is BETTER than another because they do different things and each is more appropriate to a given situation/experiment.

I do think people will take you seriously as long as you have an interesting scientific question and use an appropriate tool to answer it, the age of the tool is not important. The main issue I guess, as you pointed out, is maintenance. If anything breaks and there is no technical support, then the machine goes to the bin, which i a major incentive not to invest in it.

Roche officials told me the expiration date of all 454 reagents in the near future will all be set to December 2016. You may stockpile 454 reagents as many as you want, but there will be no technical support after that date. 

Stephane, this is just my opinion, but I also think the days of (relatively) small individual labs having their own machine is rapidly coming to an end (or already has done so).  The cost of sequencers is high, and it is very difficult to justify them (in a return on investment argument) unless you are using them a lot (as in literally 24/7).  There is also still a non-trivial skill required to get high quality NGS data out, and that means retaining skilled technicians which is an additional expense.  If the sequencer is not used frequently, then those techs also never really get to hone their skill set fully.

Those facts alone moves things practically into the realm of large sequencing cores and outright commercial sequencing centers.  They can afford the investment (equipment and people), and by combining internal users and outside contracted users, they can keep the machines active enough to provide a reasonable return on investment (especially these days when the life span of a given NGS technology is measured in, at best, a couple or few years).

Smaller individual labs are then much better off having their sequencing done by their institution or company's core facility, or contracting it out to a commercial center.  You can of course receive nothing but the raw read data back and do all the QC and data analysis yourself.  And, by not having incurred the expense of establishing and maintaining a sequencer yourself, you can invest in skilled and knowledge people to organize, analyze and interpret the actual data instead of merely generating it.

Sequencing these days, on a per-base or per-sample basis, is not a terribly expensive effort if you do not have to bear the full cost of the equipment purchase, depreciation, maintenance and operation, including the manpower costs.  But, if you are doing it entirely in-house, and not able to run samples constantly over the several year life span of a given technology, then the actual incurred costs go up substantially to the point it simple is not cost effective at all anymore.

PACBIO is the new standard for NGS of long reads Stephane! You'll get a hell of a lot more sequence than 454 ever gave 

Russell is correct. We can now sequence 2Kb amplicons on the PacBio with good accuracy because the N50 raw  sequence length is now >9kb. It means you can sequence 3-5 times the same 2Kb molecule and build a consensus sequence from that. The cost per base is also lower than on the 454, and without the higher error rates in low complexity regions.

On the lower end, you can do a Paired end 300bp run on the Illumina MiSeq to sequence 500-550 bp long DNA fragments. Since the paired sequences overlap by 50bp or more, you can join them and get 25 million 500-550bp sequences for much cheaper then the 454.

Roche is losing money on the 454 because it is an outdated technology

PACBIO gives less reads then the 454 and is very expensive in use and not suitable for every purpose, like a mix of long range PCR products for instance.

At this moment the 454 is still the best for long reads 500 - 800 bp or longer if you have experience with this machine. We have 5 years experience on this platform.

I agree with others, that you need to be skilled to get good results and this takes time, maybe to long if you start just now. But for other platforms you need to be skilled as well, to obtain long reads, so the 454 might be a good learning platform, for the near future , to switch after some time to PACBIO or Ion Torrent, which will produce longer reads at that time for less money.

Thank you all for your responses. I don't know much about PACBIO's latest specs but will definitely look into it.

Just a comment on Michael's reply about small labs outsourcing their NGS analyses. This is pretty much what I have been doing and I agree with you that most labs do that now. But I still like to be able to tweak and test the protocols before ordering large (and expensive) NGS analyses. I hope there will still be more room in the future to do that. One example could be the MinIon which is small, (supposedly cheap) and even 'disposable'. 

"But I think in terms of sequencing technology it is difficult to say that one is BETTER than another because they do different things and each is more appropriate to a given situation/experiment."

Actually, no. As others have pointed out, and as has been shown in countless publications, PacBio outperforms 454 in all qualitative and quantitative measurements that count.

How long the reagents can be stored should not be your only concern. 

We've had a 454 in our lab since 2010 and sadly, it needs frequent service by a trained technician.  Once Roche stops supporting it, I suspect within a year our instrument will become useless.  

I also agree with other posts that the learning curve for NGS is steep and there is a high cost associated with learning the lengthy and complicated 454 FLX protocols. Other platforms are much easier. 

PacBio does NOT outperform 454 in throughput. You get far fewer individual reads (not counting multiple reads of the same molecule) in PacBio. My lab definitely would like to have 454 around for our applications and we have tried PacBio.

Does anyone know which facility can still provide 454 pyrosequencing service (GS Junior system is fine) in US? Both facilities we sent our library for analysis were close : ( 

We recently sent samples to AGTC laboratory at the University of Kentucky. I would imagine they are still providing this service. 

 Stephane, thank you very much for your information