Nat. Methods: (Next)^2 Generation Sequencing Roundup

[ECO]

Nature Methods has published a short article (here) reviewing the current state of "next-next" (third?) generation sequencing technologies.

The main points they touch on include:

• The Heliscope from Helicos, no real new information, just about how advanced single molecule systems are. Oh and that you'll need 14 TB of storage per run!

• Visigen's FRET-based technology, which relies on FRET between fluorophores on the polymerase and dNTP to produce a real-time "movie" of base incorporation from a single reagent injection. Pretty awesome stuff.

• Steven Block's (Stanford) method of sequencing via "nanomechanical" properties of DNA-polymerase interactions. His group has built sensors that can detect the motion of a single protein molecule; in this case it is used to measure the movement of a polymerase along a template. By limiting nucleotides and calculating based on where the enzyme pauses, the sequence can be ascertained.

• New York-based Reveo has developed a "nano-knife edge probe" which basically drags across the DNA seeking a specific electrical signal from a specific base. (Remember, I'm a biologist, not a physicist! ) Multiple parallel probes can determine all bases.

• The usual short mention of nanopore sequencing, which is similar to the previous method but the pores are stationary and the DNA moves through the nanopore for probing.

Interesting stuff! Thanks to sci_guy for pointing it out.

[Malarky67]

real time base calling

I guess at present-- storage cost ~= reagents, though in 2-3 years time a TB maybe a bit cheaper.

However will everybody want to store that raw data. I think the problem is that you need to store the data as it is generated to analyse the base calls at the end. However if you could analyse the base calls and quality scores as the data was being generated you maybe not care about storing all raw data.

Its just a thought -- and I have no idea if it is possible.

[apfejes]

Depends what you mean by raw data. For Illumina runs, we throw away the images after they've been processed, but preserve the prb and seq (probability/quality and called sequence files.) You never know when you'll need to re-analyze or re-align a set of sequences.

Still, the size of prb and seq files isn't that great compared to the image files that are produced by the machine. (Last time I checked, about a year ago, the Illumina machine produces pretty inefficiently encoded images.)