First (?) Ion Torrent Paper out today

[lek2k]

http://www.nature.com/nature/journal...ture10242.html

Most have probably read this
The blog post on Genetic Future discusses the accuracy compared to the SoLiD.

http://www.wired.com/wiredscience/2011/07/how-accurate-is-the-new-ion-torrent-genome-really/

[nickloman]

Two other papers with Ion Torrent data (both E. coli)

BGI/Muenster http://www.nejm.org/doi/full/10.1056...=featured_home (includes the MIRA assembly I did back on 2nd June - http://pathogenomics.bham.ac.uk/blog...nome-assembly/)

Life Tech http://www.plosone.org/article/info:...l.pone.0022751

[steinmann]

Very interesting. Especially Supplementary figure 6 and 7.

They propose that they could scale up their chips to 1.1 billion sensors (0.5 um wells), which would enable high coverage human genome sequencing in a single run.

I do not doubt that they will be able to to manufacture the chips, but what I am not so sure about is how that would impact data quality. They would be forced to go down to ~0.4 um beads for their proposed 0.5 um well chip. Signal strength and signal to noise ratio are a function of how many amplicons fit on the bead surface. Their data in figure S6 indicates that going from 2 um to 1 um beads (~ 4 fold reduction in surface) already appears worsens the signal to noise ratio considerably. Going down to 0.4 um would mean reduction of bead surface of 25 fold from the original design that has a signal to noise ratio of 10.

Could anyone here speculate on how badly the this reduction in bead size and therefore signal strength would impact the signal to noise ratio and base call quality?

[Michael.James.Clark]

Oooo. Congrats, Nils!

[nilshomer]

And congrats to Nils on the authorship... !!

Thanks my friend, .

[srasdk]

I am interested to see what the error rates are like for the data presented by ion.

If you interested to sample Ion Torrent data, check:

http://www.ncbi.nlm.nih.gov/sra/SRX067313

http://www.ncbi.nlm.nih.gov/sra/SRX072974

[hyjkim]

It's a pity that their data is not accessible on the SRA. I am interested to see what the error rates are like for the data presented by ion.

From what I read in the article, Ion is not ready for human genome sequencing. at 2 hours/per * 2,000 runs for 10x coverage would require 166 days of constant sequencing with no breaks.

[jason_a]

i think it's safe to call it the first Ion Torrent paper. the first paper with PGM data that i know of is the one from the Schuster group where they used PGM for variant validation.

http://www.ncbi.nlm.nih.gov/pubmed/21709235

http://www.pnas.org/content/early/20...102838108.long

[ECO]

And congrats to Nils on the authorship... !!

[ECO]

http://www.nature.com/nature/journal/v475/n7356/full/nature10242.html