SEQanswers

Go Back   SEQanswers > Sequencing Technologies/Companies > The Pipeline



Similar Threads
Thread Thread Starter Forum Replies Last Post
"allele balance ratio" and "quality by depth" in VCF files efoss Bioinformatics 2 10-25-2011 11:13 AM
Relatively large proportion of "LOWDATA", "FAIL" of FPKM_status running cufflink ruben6um Bioinformatics 3 10-12-2011 12:39 AM
The position file formats ".clocs" and "_pos.txt"? Ist there any difference? elgor Illumina/Solexa 0 06-27-2011 07:55 AM
"Systems biology and administration" & "Genome generation: no engineering allowed" seb567 Bioinformatics 0 05-25-2010 12:19 PM
SEQanswers second "publication": "How to map billions of short reads onto genomes" ECO Literature Watch 0 06-29-2009 11:49 PM

Reply
 
Thread Tools
Old 02-20-2012, 11:29 PM   #41
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 750
Default

Quote:
Originally Posted by pmiguel View Post
Strange, does Rothberg think the genome size of phage lambda is 5000 bases? 5x10^3 to 6x10^9 that would be six orders of magnitude. Lambda is about 48,500 bp.

Okay, I am being pedantic. It does not change his main point. Whether 5 orders of magnitude or 6 that does seem like a lot of ground to cover in less than a year. Why not sequence a bacterial genome? Or yeast? Would be really nice to see the sequence reads from a data set.
The actual output of the sequencers was probably a little more than 100kb. That lambda run was sequenced on a MinION (i.e. the "smaller" sequencer), generating complete 100kb fragments from single pores:

http://www.nanoporetech.com/news/press-releases/view/39

The MinION has 512 pores, and runs can take up to 6 hours, so it would be conceivable that the run that produced this sequence could have produced 10-50Mb of sequence in six hours depending on how many active sites are assumed (i.e. 2 more orders of magnitude than what is suggested in your statement). The GridION system can do 2000 pores and run for a bit longer. Assuming that's a ten times longer run time with similar processing length / quality, you get another 40~100x quantity of generated sequence. The GridION system can be racked, so rack up 10 GridIONs together and it comes to the magic 5 orders of magnitude using the systems that have already been reported in the press releases.

The press release I linked to above suggests a 20-node GridION system with 8000 pores per cartridge is feasible, and "would be expected to deliver a complete human genome in 15 minutes", and this agrees mostly with the back-of-the-computer-screen calculations I've done here.
gringer is online now   Reply With Quote
Old 02-21-2012, 03:18 AM   #42
Pongo_T
Junior Member
 
Location: Germany

Join Date: Feb 2012
Posts: 4
Default

Quote:
Originally Posted by gringer View Post
That lambda run was sequenced on a MinION (i.e. the "smaller" sequencer)
http://www.nanoporetech.com/news/press-releases/view/39
Are your quite sure it was actually done on the Minion or on any chip array? Was this made clear in the talk?
Pongo_T is offline   Reply With Quote
Old 02-21-2012, 03:28 AM   #43
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 750
Default

Quote:
Originally Posted by Pongo_T View Post
Are your quite sure it was actually done on the Minion or on any chip array? Was this made clear in the talk?
Hmm, I think I combined a few too many things because it was in a press release that had "MinION" in the title. The important thing that I was trying to get across was that this was a sequence from a single pore, and the sequencers will have hundreds or thousands of pores on their surface.

If it is assumed that it was on a different system from the MinION (e.g. on a single isolated pore) with a run time in excess of ~30h, you're welcome to take an order of magnitude off my estimates (so still one more to go, assuming a somewhat pessimistic calculation).

Last edited by gringer; 02-21-2012 at 03:36 AM.
gringer is online now   Reply With Quote
Old 02-21-2012, 03:42 AM   #44
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 750
Default

Quote:
Originally Posted by gringer View Post
If it is assumed that it was on a different system from the MinION (e.g. on a single isolated pore) with a run time in excess of ~30h, you're welcome to take an order of magnitude off my estimates (so still one more to go, assuming a somewhat pessimistic calculation).
Here's an article that mentions tens of gigabases of sequence per day on a single GridION node with 2000 pores:

http://www.azonano.com/news.aspx?newsID=24324

This works out to 5 megabases per pore per day (assuming full pore occupancy), or about 200kb per hour. At this rate, a 100kb sequence should complete in the 6-hour run time of a MinION (even if assuming a 1/10th read speed compared to the GridION cartridges), so I'm going to stick with my initial "plausible, based on press releases" thoughts.

Last edited by gringer; 02-21-2012 at 03:49 AM. Reason: that should be 200kb, not 20kb
gringer is online now   Reply With Quote
Old 02-21-2012, 04:16 AM   #45
pmiguel
Senior Member
 
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,184
Default

Quote:
Originally Posted by gringer View Post
The actual output of the sequencers was probably a little more than 100kb. That lambda run was sequenced on a MinION (i.e. the "smaller" sequencer), generating complete 100kb fragments from single pores:

http://www.nanoporetech.com/news/press-releases/view/39

The MinION has 512 pores, and runs can take up to 6 hours, so it would be conceivable that the run that produced this sequence could have produced 10-50Mb of sequence in six hours depending on how many active sites are assumed (i.e. 2 more orders of magnitude than what is suggested in your statement).
That is exactly the point. If the MinION is capable of such a feat, then why is the only report of its sequencing DNA that of single read lambda genomes? (And where are the reads generated during the run?) Why not knock off E. coli in 6 hours? Use a dam- dcm- strain if methylation is a problem initially. Also, why not release the sequence reads? If a human genome can be sequenced on a score of GridIONs in 15 minutes, why not do that?

But since our reaction has been one of nearly blanket acceptance, it looks like Oxford Nanopore made the right call. Why release data when nearly everyone will accept what you say without question?

If we were talking about a instrument a couple of years from commercial release, then I would say our attitude of polite astonishment would be warranted. But later this year?

I think two extreme possibilities include:

(1) Everything is exactly the way it is portrayed in the press releases. No data release because they know they will have real instruments available for sale in 6 months and they will sell every single one they make.

(2) There has been some sort of disconnect between what is being marketed and what is actually possible. No data is being released, because there is nothing to release.

Seems like our initial reaction was to presume possibility 1. Rothberg presumes something like 2. I don't see personally see any information that makes it possible to distinguish between the two possibilities. You can claim "sour grapes" all you want about Rothberg's comments. But at this point his claims are backed by the same amount of data that has been publicly released as Oxford Nanopore's claims: zero.

I am just wondering though, did anyone at the presentation at AGBT ask for the read data from the lambda genome sequencing?

I cannot imagine that I would have at that juncture. PacBio had been trumpeting their potential for years by the time they were 6 months from a commercial release. By that time the questions about their capabilities had become pointed. My impression is that Oxford Nanopore has taken a much more low key approach until now. We have all seen amazing events unfold in our field over the last 5 years. Nearly anything seems possible. Still, we should ask for data before we allow ourselves to become part of the hype. Has anyone?

--
Phillip
pmiguel is offline   Reply With Quote
Old 02-21-2012, 06:48 AM   #46
pmiguel
Senior Member
 
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,184
Default

Quote:
Originally Posted by gringer View Post
Here's an article that mentions tens of gigabases of sequence per day on a single GridION node with 2000 pores:

http://www.azonano.com/news.aspx?newsID=24324

This works out to 5 megabases per pore per day (assuming full pore occupancy), or about 200kb per hour. At this rate, a 100kb sequence should complete in the 6-hour run time of a MinION (even if assuming a 1/10th read speed compared to the GridION cartridges), so I'm going to stick with my initial "plausible, based on press releases" thoughts.
Actually, if you have access, the In Sequence report on the *ION is pretty comprehensive.

What I am looking for at this point is someone who has used an instrument reporting what it does right now. (This could be Oxford Nanopore.) So far, it looks like the only concrete claim of that sort is the lambda genome one. If the third hand reports are accurate in this respect. Are we absolutely sure the intent of that report was not -- "in principle, given the potential of this device, it could sequence both strands of a lambda genome in a single read"?

Yes, I realize there might be lots of data tied up by NDAs. Maybe there is high tier journal article embargo in effect that is preventing more disclosure? Oxford Nanopore may have their hands tied by British advertising regulations, not wanting to release 4% error data when they think they will hit 1% in a brief time span or some other issue. Does not matter, I think Rothberg's doubts are doubts we should all consider.

--
Phillip
pmiguel is offline   Reply With Quote
Old 02-21-2012, 09:04 AM   #47
singlemoleculer
Junior Member
 
Location: USA

Join Date: Jan 2012
Posts: 8
Default Deletion rate of 17%

Based on Oxford Nanopore's press release (http://www.nanoporetech.com/news/press-releases/view/39) and their spelling of lambda, looks like they are trying to prepare us for lots of deletions.
singlemoleculer is offline   Reply With Quote
Old 02-21-2012, 09:11 AM   #48
Nanoporous
Junior Member
 
Location: San Diego

Join Date: Dec 2009
Posts: 7
Default

Quote:
Originally Posted by Pongo_T View Post
For the more technically minded: did Clive Brown show any primary data in the sense of simultaneously acquired nanopore current traces (there is none on the webpage, is there?) It is one thing to show that (1) strand sequencing basically works on short genomes and that (2) you have a mulitplexed nanopore chip which might be good for longer ones, but quite another to show that strand sequencing has actually successfully been done using that chip. There are lots of things where they must have made tremendous progress, not least the on-chip multiplexed 512-channel (for the Minion) voltage-clamp amplifier (CMOS-based, I understand) which has to have pretty low noise, a.s.f.
He showed very little data. There was certainly no data showing simultaneous currents. He did show some current versus time traces and how they remove the time information to obtain a spatial trace. I did not see 64 different current levels being detected. This is not to say they don't have such data. The presentation was only 17minutes, and he had a lot to cover.

He did show some data on how there is no degradation of the signal over a long period - which is kind of obvious, this being electrical signal and not fluorescent probes. He also showed current traces at various frequencies, but did nto specify the noise levels. From a quick visual scaling, I thought noise was ~20pA at 32kHz.
Nanoporous is offline   Reply With Quote
Old 02-21-2012, 10:40 AM   #49
Geneus
Member
 
Location: New Jersey

Join Date: Dec 2010
Posts: 60
Default

Quote:
Originally Posted by pmiguel View Post
Actually, if you have access, the In Sequence report on the *ION is pretty comprehensive.

What I am looking for at this point is someone who has used an instrument reporting what it does right now. (This could be Oxford Nanopore.) So far, it looks like the only concrete claim of that sort is the lambda genome one. If the third hand reports are accurate in this respect. Are we absolutely sure the intent of that report was not -- "in principle, given the potential of this device, it could sequence both strands of a lambda genome in a single read"?

Yes, I realize there might be lots of data tied up by NDAs. Maybe there is high tier journal article embargo in effect that is preventing more disclosure? Oxford Nanopore may have their hands tied by British advertising regulations, not wanting to release 4% error data when they think they will hit 1% in a brief time span or some other issue. Does not matter, I think Rothberg's doubts are doubts we should all consider.

--
Phillip
My take on this...ONT is holding their cards very tight to their chest. And why not? Anticipation is a good thing from a sales and marketing perspective. I suspect they (ONT) will not disappoint when they are ready to provide the "hard" data.
Geneus is offline   Reply With Quote
Old 02-21-2012, 10:47 AM   #50
ECO
--Site Admin--
 
Location: SF Bay Area, CA, USA

Join Date: Oct 2007
Posts: 1,352
Default

Quote:
Originally Posted by pmiguel View Post
But since our reaction has been one of nearly blanket acceptance, it looks like Oxford Nanopore made the right call. Why release data when nearly everyone will accept what you say without question?
Not everyone!

Based on this page (http://www.nanoporetech.com/news/press-releases/view/20), ONT has raised a total of 74 million GBP ($120 million USD), with $41 million of it coming last April (meaning, they needed more money 10 months ago).

I would be extremely surprised if they can commercialize a single molecule sequencer that requires semiconductor manufacturing for this amount of money.
  • R&D (likely electrical/sensor engineers, protein engineering/enzymologists, crystallographers, bioinformaticians, signal processing, probably more.) I would guess >50 people.
  • Instrument design/compliance (2-3)
  • Instrument manufacturing (20 ?)
  • Chip manufacturing (team to handle outsourced mfg and incoming QC)
  • Reagent manufacturing (5)
  • Quality testing of all of the above (5)
  • Licensing fees ($$ ??)
  • Worldwide sales network (20 at least?)
  • Tech support (5)
  • Application scientists (5)
  • The rest of SG&A (10)
This is >100 people...likely what they need NOW to commercialize in the middle/end of 2012. Maybe they have it already, or maybe they are kaizen/black-belt/efficiency ninjas.


All I know is, I saw a video of a GridION over a year ago, and we still haven't seen a fastq from it. That should set off all kinds of alarm bells.


</coffee-fueled-skepticism>
ECO is offline   Reply With Quote
Old 02-21-2012, 11:15 AM   #51
aeonsim
Member
 
Location: Belgium

Join Date: Jun 2011
Posts: 45
Default

Quote:
Originally Posted by pmiguel View Post
I am just wondering though, did anyone at the presentation at AGBT ask for the read data from the lambda genome sequencing?
--
Phillip

No idea about this, but a few people have been messaging them on Twitter asking for the presentation release and some data. Might be worth trying if you can get enough people tweeting them about it we may at least get a reply.
aeonsim is offline   Reply With Quote
Old 02-21-2012, 11:15 AM   #52
pmiguel
Senior Member
 
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,184
Default

Quote:
Originally Posted by Geneus View Post
My take on this...ONT is holding their cards very tight to their chest. And why not? Anticipation is a good thing from a sales and marketing perspective. I suspect they (ONT) will not disappoint when they are ready to provide the "hard" data.
Okay that might be what they want. I say what we should want is data. Release the reads ONT!

--
Phillip
pmiguel is offline   Reply With Quote
Old 02-21-2012, 11:16 AM   #53
nickloman
Senior Member
 
Location: Birmingham, UK

Join Date: Jul 2009
Posts: 356
Default

ECO -Not sure if you can read that much into the finances.

Solexa had raised only $40m total by the end of 2004 and were well on their way to the 1G Genetic Analyzer at that point.

Bear in mind this is a British company - we make a virtue out of doing a lot for very little! Clive probably still has to bring in his own tea-bags
nickloman is offline   Reply With Quote
Old 02-21-2012, 12:11 PM   #54
ECO
--Site Admin--
 
Location: SF Bay Area, CA, USA

Join Date: Oct 2007
Posts: 1,352
Default

Quote:
Originally Posted by nickloman View Post
ECO -Not sure if you can read that much into the finances.

Solexa had raised only $40m total by the end of 2004 and were well on their way to the 1G Genetic Analyzer at that point.
It's the only data I have to "read into".

But overall, I agree. However, the technological hurdles that ONT claims to have overcome (and packaged into a USB drive) are _far_ more complicated than a microscope pointed at a slide.
ECO is offline   Reply With Quote
Old 02-21-2012, 07:00 PM   #55
ymc
Senior Member
 
Location: Hong Kong

Join Date: Mar 2010
Posts: 484
Default

Can someone tell me how to do this math??

300bp/s/pore == 8.64Gbp/hr/node

50x coverage means 150Gbp

20*8.64/4 == 43.2Gbp in 15 min with 20 nodes

150Gbp >> 43.2Gbp???

http://www.genomeweb.com/sequencing/...ing-instrument

"A 20-node installation, using 8,000-nanopore cartridges, is expected to deliver a complete human genome at 50-fold coverage in 15 minutes, according to the company, or 3 terabases of data per day, based on a sequencing speed of 300 bases per second. For that setup, the cost per gigabase is expected to be under $10."
ymc is offline   Reply With Quote
Old 02-21-2012, 10:59 PM   #56
Wallysb01
Senior Member
 
Location: San Francisco, CA

Join Date: Feb 2011
Posts: 286
Default

Quote:
Originally Posted by Nanoporous View Post
He did show some data on how there is no degradation of the signal over a long period - which is kind of obvious, this being electrical signal and not fluorescent probes. He also showed current traces at various frequencies, but did nto specify the noise levels. From a quick visual scaling, I thought noise was ~20pA at 32kHz.
Ion Torrent doesn't use fluorescence and they have problems maintaining quality as reads get longer. And if we're talking about 10Kb+, it doesn't take much degradation rate per base to be totally useless. Then of course comes the question of, what does the quality start at? Maybe you don't lose much over time, but if you don't start very high, who cares?

Last edited by Wallysb01; 02-21-2012 at 11:03 PM.
Wallysb01 is offline   Reply With Quote
Old 02-22-2012, 03:36 AM   #57
pmiguel
Senior Member
 
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,184
Default

Quote:
Originally Posted by Wallysb01 View Post
Ion Torrent doesn't use fluorescence and they have problems maintaining quality as reads get longer. And if we're talking about 10Kb+, it doesn't take much degradation rate per base to be totally useless. Then of course comes the question of, what does the quality start at? Maybe you don't lose much over time, but if you don't start very high, who cares?
Okay, but the Ion Torrent is a second generation instrument. That means ensembles of nascent strands reporting as each base (or series of homopolymeric bases) is incorporated. That is inherently limited by your ability to keep all the nascent strands in sync. As they begin to differ in the base they report, signal descends into noise. How the addition of a base is reported -- via pH change or fluorescence -- does not alter that limitation.

4th gen means you don't rely on synthesis of a nascent strand and you query only one strand at a time. Hence the syncing issue disappears. To be replaced with other issues, no doubt. But, to the extent you can keep that strand moving through the pore, and the pore characteristics do not change randomly over time, the read quality would not be expected to diminish as read length increases.

In principle anyway. In actual practice? Well, that is what read data would help you determine. These are extraordinary claims here. Yet not even one data set has been publicly released? Why not? If it is because there is little to gain from ONT doing so, then part of the blame is on us.

Where is the skepticism? Sure we see a little in this forum. But in the media at large I see only Rothberg's comments -- and his comments are being largely discounted because of his role in a competing instrument system.

--
Phillip
pmiguel is offline   Reply With Quote
Old 02-22-2012, 06:13 AM   #58
GW_OK
Senior Member
 
Location: Oklahoma

Join Date: Sep 2009
Posts: 382
Default

Just a quick question on generations. We're all aware of the 2nd gen stuff we've got in our labs, but people are calling ONT 4th or 5th gen. What was 3rd gen? Pacbio?
GW_OK is offline   Reply With Quote
Old 02-22-2012, 08:39 AM   #59
krobison
Senior Member
 
Location: Boston area

Join Date: Nov 2007
Posts: 747
Default

Quote:
Originally Posted by pmiguel View Post

I think two extreme possibilities include:

(1) Everything is exactly the way it is portrayed in the press releases. No data release because they know they will have real instruments available for sale in 6 months and they will sell every single one they make.

(2) There has been some sort of disconnect between what is being marketed and what is actually possible. No data is being released, because there is nothing to release.

Seems like our initial reaction was to presume possibility 1. Rothberg presumes something like 2. I don't see personally see any information that makes it possible to distinguish between the two possibilities. You can claim "sour grapes" all you want about Rothberg's comments. But at this point his claims are backed by the same amount of data that has been publicly released as Oxford Nanopore's claims: zero.

--
Phillip
Of course, the funny part about Rothberg making these comments is it wasn't too many AGBTs ago that the exact same list of complaints was being leveled at him after his presentation similarly electrified the sequencing community (or should I say ionized?)
krobison is offline   Reply With Quote
Old 02-22-2012, 08:53 AM   #60
Nanoporous
Junior Member
 
Location: San Diego

Join Date: Dec 2009
Posts: 7
Default

Quote:
Originally Posted by krobison View Post
Of course, the funny part about Rothberg making these comments is it wasn't too many AGBTs ago that the exact same list of complaints was being leveled at him after his presentation similarly electrified the sequencing community (or should I say ionized?)
Rothberg's comments are perhaps a bit extreme, and always comes with the 'he's a competitor' tag, but after the initial euphoria, lot of people at AGBT associated with both genome sequencing technology and genomics had some reservations. The general mood was 'this is a great technology, but show me the real data'.
Nanoporous is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 03:11 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO