Comments on: The Future of Second Generation Sequencing http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/ Sun, 22 Jan 2012 22:30:33 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Science Report » Blog Archive » Nibbles: Sequencing, Agricultural origins, Mating systems, Tomato shelf-life, Beer vs Tea, Soy, Carrot, Seed processing, Screw-pine, Yams, Salicornia, Pollinators http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-2050 Science Report » Blog Archive » Nibbles: Sequencing, Agricultural origins, Mating systems, Tomato shelf-life, Beer vs Tea, Soy, Carrot, Seed processing, Screw-pine, Yams, Salicornia, Pollinators Mon, 08 Feb 2010 12:46:01 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-2050 [...] Second generation sequencing on the one hand. First generation methylation mapping on the other. What’s a poor bitechnologist to do? [...] [...] Second generation sequencing on the one hand. First generation methylation mapping on the other. What’s a poor bitechnologist to do? [...]

]]> By: Nibbles: Sequencing, Agricultural origins, Mating systems, Tomato shelf-life, Beer vs Tea, Soy, Carrot, Seed processing, Screw-pine, Yams, Salicornia http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-2049 Nibbles: Sequencing, Agricultural origins, Mating systems, Tomato shelf-life, Beer vs Tea, Soy, Carrot, Seed processing, Screw-pine, Yams, Salicornia Mon, 08 Feb 2010 10:27:54 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-2049 [...] Second generation sequencing on the one hand. First generation methylation mapping on the other. What’s a poor bitechnologist to do? [...] [...] Second generation sequencing on the one hand. First generation methylation mapping on the other. What’s a poor bitechnologist to do? [...]

]]> By: Luke http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-1710 Luke Wed, 13 Jan 2010 23:50:36 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-1710 Yeah Complete Genomics is definitely worth keeping an eye on. As well as the obvious fact that their business model is very different, I also know that they are working on some interesting chemistry-side (as opposed to software-side) technologies to make mapping and assembly much more accurate, which if they get working well could overcome many of the disadvantages of the 2nd gen technology. If I was going to place a bet on who wins the Archon X prize, it'd be Complete Genomics. Yeah Complete Genomics is definitely worth keeping an eye on. As well as the obvious fact that their business model is very different, I also know that they are working on some interesting chemistry-side (as opposed to software-side) technologies to make mapping and assembly much more accurate, which if they get working well could overcome many of the disadvantages of the 2nd gen technology.

If I was going to place a bet on who wins the Archon X prize, it’d be Complete Genomics.

]]> By: Marco Dotto http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-1709 Marco Dotto Wed, 13 Jan 2010 22:50:10 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-1709 What about Complete Genomics? ...it looks like it's in the mid way between 2nd and 3rd gen sequencing techs. They talked about 5k human genomes delivered in 2010. And in their proof of principle paper they were talking about 8k $ for a 87X coverage (compared to 10K $ for 30X with the new Illumina HiSeq 2000). What do you think about Complete genomics? Their strategy is really new for this market since they are not going to sell the sequencers but only the sequencing service and they seem to be ready to penetrate the market starting from 2011 (since the genomes for 2010 are already sold to . Their error rate seems very good too: accuracy of 1 false variant per 100 kilobases. What about Complete Genomics? …it looks like it’s in the mid way between 2nd and 3rd gen sequencing techs. They talked about 5k human genomes delivered in 2010. And in their proof of principle paper they were talking about 8k $ for a 87X coverage (compared to 10K $ for 30X with the new Illumina HiSeq 2000).
What do you think about Complete genomics? Their strategy is really new for this market since they are not going to sell the sequencers but only the sequencing service and they seem to be ready to penetrate the market starting from 2011 (since the genomes for 2010 are already sold to . Their error rate seems very good too: accuracy of 1 false variant per 100 kilobases.

]]> By: Luke http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-1707 Luke Wed, 13 Jan 2010 22:01:44 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-1707 24 Gb per day is the maximum scanning rate of the Helicos imaging system. According to the system specs (which are a year out of date), the throughput is 2.5-3.4 Gb per day (105-140 Mb/hour); i.e. slightly less than the GAII, and much less than HiSeq. The error rates of Helicos are a bit of a worry too - their whole-genome paper last year reported error rates of 3.6%, and most of these were deletions, making indel detection difficult. It is always possible that Helicos will make some kind of surprise breakthrough, but with such high error rates and such low read length, the machines suffer from 2nd-gen style problems even more than 2nd gen machines. Given the recent <a href="http://www.genomeweb.com/compugen-shares-soar-december-helicos-falls" rel="nofollow">fall in share price</a>, I don't think the current Heliscope is really going to play much of a part in the future of sequencing. In terms of what system has the most room to grow; that really depends on what time scale you are talking about. In the short term, I expect Illumina still has a lot of development in store; the R&D guys have been doing this a long time, and have an awful lot of support from the company. By 2013/14, I expect the competition will be between Oxford NanoPore and PacBio for who can push their read length furthest into the tens of thousands. 24 Gb per day is the maximum scanning rate of the Helicos imaging system. According to the system specs (which are a year out of date), the throughput is 2.5-3.4 Gb per day (105-140 Mb/hour); i.e. slightly less than the GAII, and much less than HiSeq.

The error rates of Helicos are a bit of a worry too – their whole-genome paper last year reported error rates of 3.6%, and most of these were deletions, making indel detection difficult.

It is always possible that Helicos will make some kind of surprise breakthrough, but with such high error rates and such low read length, the machines suffer from 2nd-gen style problems even more than 2nd gen machines. Given the recent fall in share price, I don’t think the current Heliscope is really going to play much of a part in the future of sequencing.

In terms of what system has the most room to grow; that really depends on what time scale you are talking about. In the short term, I expect Illumina still has a lot of development in store; the R&D guys have been doing this a long time, and have an awful lot of support from the company. By 2013/14, I expect the competition will be between Oxford NanoPore and PacBio for who can push their read length furthest into the tens of thousands.

]]> By: Todd http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-1705 Todd Wed, 13 Jan 2010 20:54:45 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-1705 Helicos claims at least 24Gbase per 24Hrs and 36Gbase per 1.5days which is the exact same throughput claimed for HiSeq2000. http://helicosbio.com/Technology/TrueSingleMoleculeSequencing/tSMStradePerformance/tabid/151/Default.aspx Which platform do you suspect has the most room to grow? Which platform has the lowest error rate? HLCS needs to get their system cost down to $600K or so to be competitive (they are working on this). Helicos claims at least 24Gbase per 24Hrs and 36Gbase per 1.5days which is the exact same throughput claimed for HiSeq2000.

http://helicosbio.com/Technology/TrueSingleMoleculeSequencing/tSMStradePerformance/tabid/151/Default.aspx

Which platform do you suspect has the most room to grow?
Which platform has the lowest error rate?

HLCS needs to get their system cost down to $600K or so to be competitive (they are working on this).

]]> By: Another Stop on the Road to the $1,000 Genome http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-1703 Another Stop on the Road to the $1,000 Genome Wed, 13 Jan 2010 19:58:15 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-1703 [...] 1/13: Daniel MacArthur of Genetic Future, Luke Jostins of Genetic Interference and David Dooling of PolITiGenomics all offer their analysis of what Illumina’s announcement [...] [...] 1/13: Daniel MacArthur of Genetic Future, Luke Jostins of Genetic Interference and David Dooling of PolITiGenomics all offer their analysis of what Illumina’s announcement [...]

]]> By: Daniel MacArthur http://www.genetic-inference.co.uk/blog/2010/01/the-future-of-second-generation-sequencing/comment-page-1/#comment-1698 Daniel MacArthur Wed, 13 Jan 2010 14:47:04 +0000 http://www.genetic-inference.co.uk/blog/?p=752#comment-1698 Nicely put, Luke. I'm excited about the launch of third-gen technologies this year simply because I'm desperately craving the end of the informatic nightmare that is short-read sequencing; but clearly no-one will be retiring their GAs and HiSeqs in favour of third-gen platforms in the immediate future. It will certainly be interesting to see how quickly third-gen gets taken up as a supplement to second-gen for human WGS: imagine hybrid de novo assemblies combining low-coverage third-gen data with high-coverage second-gen sequence. Sweet. Nicely put, Luke. I’m excited about the launch of third-gen technologies this year simply because I’m desperately craving the end of the informatic nightmare that is short-read sequencing; but clearly no-one will be retiring their GAs and HiSeqs in favour of third-gen platforms in the immediate future.

It will certainly be interesting to see how quickly third-gen gets taken up as a supplement to second-gen for human WGS: imagine hybrid de novo assemblies combining low-coverage third-gen data with high-coverage second-gen sequence. Sweet.

]]>