Saturday, February 09, 2008

Marco Island - Saving Some of the Best for Last

Well, the Marco Island AGBT meeting just wrapped up and I they definitely saved some of the best for the end. I do not have a ton of time but here are my favorites:

Stephan Schuster gave a funny, entertaining and interesting talk on mammoth mitochondrial genomics. He riddled the talk with funny stories and one liners about his efforts to sequence old DNA and to publish the results. He did a good job of showing why Roche-454 sequencing is quite ideally suited for studies of old DNA.

John Leamon from Raindance Technologies summarize some of their work on droplet based microfluidics. He showed videos of droplets moving through their system and showed how it sould be used for various digital PCR-like activities.

But it was the last talk of the whole meeting that really did blow my mind. It was from Steve Turner from Pacific Biosciences. He presented an overview of their sequencing technology as well as a tiny bit of data. Now, normally I am uninterested in marketing talks where little data is presented. But this talk was different. First, their technology clearly has enormous potential for revolutionizing the sequencing field. Basically, what they are doing is reading the activity of a DNA polymerase as it replicates a single DNA molecule and they do it in real time. He referred to this as using the DNA polymerase as a sequencing engine and then he took the crowd through the details of the technology and some of the modifications they have made to make it work better. I will try to post later with more detail on their methods.

No - they are not quite ready for prime time yet. But the potential is pretty absurd. I see two key major advantages of their method if it can be fine tuned to work well -- (1) it is screamingly fast - because they let the polymerase do all the work in essence (2) it can potentially get long reads -- right now he claimed they could do up to 1500 base pairs and theoretically it could go much higher. If they can get this to work with reads of 10,000 bases for example, this will completely reconfigure the field. No longer will one have to worry about the complexities of mapping short reads as with many of the current "new" methods. And the long reads, coupled with many molecules per run, plus the high speed, this technology is the first I have seen that has shown some results and that could really lead to the $1000 human genome. Again, not clear when/if they will be ready for release to the world so don't hold off buying one of the other systems that currently work (i.e., Illumina, Roche, ABI Solid) if you want to do "next gen" sequencing. But this company is one to keep an eye on.



  1. it was great to meet & hang with you at ye olde AGBT meeting & plan our new company. you've got my contact info now so don't hesitate to drop a line. keep up the good work with the blog. cheers, j.

  2. Wow. Thanks for the heads-up on the technology. If this PacBio system pans out, it will change everything. If we think that bioinformatics is the bottle-neck now, with this technology, it's gonna get alot worse before it gets better.

  3. grass skirt manb --- will be in touch about the new company

    John L ... actually what I find most iimpressive about the Pacific Biosciences system is their potential read length. I think the biggest informatics issue in much of the new data coming out is the short reads ... once reads get to be 1500 bp or longer, we can stop worrying about some of the mapping/assembly problems we have with current methods.

  4. Yes, the read-length issues with assembly are indeed serious bioinformatic hurdles that may be obviated by the PacBio method. However, what I was mainly referring to is the data spigot that will be on full-blast (pun not intended, but it works) once this technology is used for de novo sequencing. It's a really exciting prospect, but many of us are already wallowing in data (from 10s to 100s of genomes). It's daunting to imagine dealing with an order of magnitude more. But it's certainly exciting to consider!

  5. I do not think it is that daunting. All we will have to learn to do is to ignore some data. Right now, with the short reads, it is even hard to figure out what to ignore. With larger reads, I think people will begin to figure out more how to ignore things. Just like most researchers do not analyze ALL rRNA sequences in genbank at once, nor do the look at all gel images in pubmed, we will stop trying to look at all sequence data at once.

  6. Did they say how far away in the future the next next generation sequencers were from being on the market (or when we might be expecting a public offering?).

    other random note, where can I get a "What would Jesus sequence" shirt?

  7. PacBio said they estimate 2010 for sequencers to be available.

    As for the shirts, I will announce this later but you can get them at

  8. Just randomly found your site - originally from Roseville, CA, I am the biologist for the City of Marco Island and my sister rcvd her PhD from UCD; she is a prof now at Pepperdine. Just an acknowledgement due to coincidences! Hope you enjoyed the Marco Island environment!

  9. Marco Island was great .... you should drop by next years genome meeting in February ...


Most recent post

My Ode to Yolo Bypass

Gave my 1st ever talk about Yolo Bypass and my 1st ever talk about Nature Photography. Here it is ...