THE FUTURE IS NOW MAKE YOUR TIME.

[Kai]

http://www.newscientist.com/article/dn21495-usb-stick-can-sequence-dna-in-seconds.html

a little gadget that can sequence DNA while plugged into your laptop

the DNA does not need to be amplified

can sequence DNA strands as long as 10,000 bases continuously

the MinION would take about 6 hours to complete a human genome

Each unit is expected to cost $900 when it goes on sale later this year

WhattheIdon'teven. No PCR, no shotgun sequencing, speed comparable to pirosequencing, fits in the palm of your hand, and COSTS LESS THAN 1000 DOLLARS.

How it works -

Oxford Nanopore is also building a larger device, GridION, for lab use. Both GridION and MinION operate using the same technology: DNA is added to a solution containing enzymes that bind to the end of each strand. When a current is applied across the solution these enzymes and DNA are drawn to hundreds of wells in a membrane at the bottom of the solution, each just 10 micrometres in diameter.

Within each well is a modified version of the protein alpha hemolysin (AHL), which has a hollow tube just 10 nanometres wide at its core. As the DNA is drawn to the pore the enzyme attaches itself to the AHL and begins to unzip the DNA, threading one strand of the double helix through the pore. The unique electrical characteristics of each base disrupt the current flowing through each pore, enough to determine which of the four bases is passing through it. Each disruption is read by the device, like a tickertape reader.

This is science fiction territory, people. Combine one of these with an iphone, and you have damn near a tricorder.


My only question in this is "are the MinIONs one time use or mult-use?" Because if they are multi-use /I. Want. One./

[Xooxe]

That's cool as hell.

I'd like to know more about why they used AHL. Like, did they go "hey look, that tubey thing that MRSA uses to breach cell membranes, and eventually pop them to feast on their iron would do the trick"?

OK, should have googled. It seems to be fairly established as a technology. http://www.ks.uiuc.edu/Research/hemolysin/

OK, not so well established. WATCH ME FUMBLE WITH GOOGLE ALL NIGHT. 15 years of research.

[Elder Iptuous]

according to the Gizmag article i saw on this, it's a 'disposable' unit.
it also says we can expect the price to drop significantly as production gins up.

[Kai]

according to the Gizmag article i saw on this, it's a 'disposable' unit.
it also says we can expect the price to drop significantly as production gins up.

Yeah. The DamION seems to be multi use though. Still probably much smaller and cheaper than any pirosequencer out there.

For those of you who are not familiar, there are 3 (now 4) generations of sequencing technology.

The first is Sanger Sequencing, which uses a process called a Polymerase Chain Reaction (or PCR for short). PCR was a radical discovery (apparently discovered after an acid trip) by Kary Mullis.

The general idea is this: you have one DNA strand and you want a whole bunch of copies. Now, you know that when you heat DNA, the double helix pulls apart into the two complimentary strands, and you know that if you added a DNA polymerase (a protein that finds single strands and builds the compliment to them) You will get two double stranded DNA helixes. Now, the problem is most DNA polymerase doesn't like getting heated, it tends to denature. So Mullis looked for bacteria in thermal hot springs and used the DNA polymerase from those. Suddenly, you could add this "taq polymerase" protein which doesn't denature under high heat to the mix, add a primer which will attach to your gene of interest, run the mixture through a successive hot-warm-hot sequence of water baths, and come out with a huge amount of DNA. Every time it goes in the hot water bath, it denatures, every time it goes in the warm water bath, the taq polymerase makes a complimentary strand.

So, now you have a whole lot of DNA. But the DNA isn't the entire length of the gene, because you've added base pairs to the solution that are a little broken, and these will randomly be used to cap the length. Which means you have a whole bunch of different lengths of DNA. And if you include only one type of broken base pair (say, a G (guanine)) in the mix, then all the lengths will be all capped at places where a Guanine would attach. Do this with the other three basepairs, and now you can place each basepair cap in it's own row on a gel electrophoresis setup, and the electric dipoles will pull the shorter strands faster than the longer ones. Basically, you'll have a visual matrix of the sequence, with each basepair in the position by length of the strand. This was updated from the gel setup into capillary tubes, but it still is rather low tech, and requires a huge amount of space to do a goodly amount of sequence. It takes years to sequence a human genome this way. Incidently, this is the method that was used for the Human Genome Project.


Second generation sequencing is somewhat the same, but much faster. It still uses PCR to amplify the genes, and still uses these broken basepairs, but instead of the kind previously, it uses ones that have a dye attached. This is called Dye Terminator Sequencing. Wash away the other basepairs, and you can clearly see the color. Now use an enzyme to cut off the dye terminator, and add another dye labeled basepair. Rinse (literally), repeat, and by the sequence of colors, you will have the sequence of the DNA. But the major problem with this is that, while it's faster than Sanger sequencing, it still takes a long time. You can do this in high thoroughput microarrays, but it still is time consuming. It also can't sequence long segments of DNA.

The solution to this has been something called shotgun sequencing, where the DNA is cut up into a whole bunch of manageable bits, and then later reassembled by software. This is the method that was used by Craig Venter in his ocean water sample sequencing work.


Third generation is called Pirosequencing, where instead of a dye, a light of different frequencies is detected whenever the next basepair is added (this is simplifying a bit; there are a bunch of other molecules involved). That light is different for the individual basepairs, so the output is a graph of the sequence order by different wavelengths of light. This is faster, but it still requires PCR, and it makes shorter sequence lengths than even first generation Sanger sequencing. But with shotgun sequencing you only get short fragments anyway. When I was in grad school, this was the method that made you drool. Third gen was wild stuff.



Fourth generation includes stuff like the nanopore technology. It's fast, it's cheap, and it can sequence very long strands of DNA without cutting them up first. PCR and shotgun sequencing no longer needed to map an entire genome. I hope now that I've described the above methods, you'll understand why my mind is blown. Compared to 4th gen, 1st gen is like banging rocks together. And it is /still/ amazing. You could still, for example, use PCR to amplyfy a particular gene in microarray, with each well being a different sample, and use nanopore electrosequencing to sequence them all at once, very quickly. Mitochondrial genomes are becoming the standard for molecular identification of animal species, for example, and you could sequence a whole bunch of those in no time at all through this method.

[Elder Iptuous]

Awesome post, Kai! 
it really puts the advances in perspective.
if you were to speculate about the possible repercussions of being able to nigh-instantly and very cheaply sequence an individual entity (with particular interest in humans), what would you see as the most radical of possibilities?
how might medicine most effectively be able to use the wealth of data if the entire population had their genome sequenced?

[Kai]

Awesome post, Kai! 
it really puts the advances in perspective.
if you were to speculate about the possible repercussions of being able to nigh-instantly and very cheaply sequence an individual entity (with particular interest in humans), what would you see as the most radical of possibilities?
how might medicine most effectively be able to use the wealth of data if the entire population had their genome sequenced?

If you had the entire genomes of a large portion of the population, and could correlate genetic diseases and disorders to these, the causes of some of these would become quickly obvious. You could determine, from birth, what medical issues a person may be subceptable to later in life. You could tailor treatment for individuals. On the pathogen side, diagnosis would become very easy.

These are of course, conservative estimates.

In the realm of radical speculation, this may very well be the step needed before people start custom wetware augmentation. I'm talking gene therapy, laboratory organ growth, part replacement and enhancement, grafting...I mean, if you can figure out how to screw with human development, turning cells back into pluripotent stem cells and guiding them through tissue growth, you could do practically anything. Want vision as good as a hawk? Or smell as good as a bloodhound? But lets consider this: when real life furries are walking around, how weird is queer going to be?

[Elder Iptuous]

the conservative estimates are incredible.  i'm curious what sort of effort would be required to implement them.  is the computation to correlate this volume of data in place already, or would it require additional advancement?
how could this be organized?  is there a medical authority that could handle the task?
it seems the possibilities are a shining jewel.  an irresistible lure.
i'm game.  let's do it!

[Mesozoic Mister Nigel]

according to the Gizmag article i saw on this, it's a 'disposable' unit.
it also says we can expect the price to drop significantly as production gins up.

Yeah. The DamION seems to be multi use though. Still probably much smaller and cheaper than any pirosequencer out there.

For those of you who are not familiar, there are 3 (now 4) generations of sequencing technology.

The first is Sanger Sequencing, which uses a process called a Polymerase Chain Reaction (or PCR for short). PCR was a radical discovery (apparently discovered after an acid trip) by Kary Mullis.

The general idea is this: you have one DNA strand and you want a whole bunch of copies. Now, you know that when you heat DNA, the double helix pulls apart into the two complimentary strands, and you know that if you added a DNA polymerase (a protein that finds single strands and builds the compliment to them) You will get two double stranded DNA helixes. Now, the problem is most DNA polymerase doesn't like getting heated, it tends to denature. So Mullis looked for bacteria in thermal hot springs and used the DNA polymerase from those. Suddenly, you could add this "taq polymerase" protein which doesn't denature under high heat to the mix, add a primer which will attach to your gene of interest, run the mixture through a successive hot-warm-hot sequence of water baths, and come out with a huge amount of DNA. Every time it goes in the hot water bath, it denatures, every time it goes in the warm water bath, the taq polymerase makes a complimentary strand.

So, now you have a whole lot of DNA. But the DNA isn't the entire length of the gene, because you've added base pairs to the solution that are a little broken, and these will randomly be used to cap the length. Which means you have a whole bunch of different lengths of DNA. And if you include only one type of broken base pair (say, a G (guanine)) in the mix, then all the lengths will be all capped at places where a Guanine would attach. Do this with the other three basepairs, and now you can place each basepair cap in it's own row on a gel electrophoresis setup, and the electric dipoles will pull the shorter strands faster than the longer ones. Basically, you'll have a visual matrix of the sequence, with each basepair in the position by length of the strand. This was updated from the gel setup into capillary tubes, but it still is rather low tech, and requires a huge amount of space to do a goodly amount of sequence. It takes years to sequence a human genome this way. Incidently, this is the method that was used for the Human Genome Project.


Second generation sequencing is somewhat the same, but much faster. It still uses PCR to amplify the genes, and still uses these broken basepairs, but instead of the kind previously, it uses ones that have a dye attached. This is called Dye Terminator Sequencing. Wash away the other basepairs, and you can clearly see the color. Now use an enzyme to cut off the dye terminator, and add another dye labeled basepair. Rinse (literally), repeat, and by the sequence of colors, you will have the sequence of the DNA. But the major problem with this is that, while it's faster than Sanger sequencing, it still takes a long time. You can do this in high thoroughput microarrays, but it still is time consuming. It also can't sequence long segments of DNA.

Fourth generation includes stuff like the nanopore technology. It's fast, it's cheap, and it can sequence very long strands of DNA without cutting them up first. PCR and shotgun sequencing no longer needed to map an entire genome. I hope now that I've described the above methods, you'll understand why my mind is blown. Compared to 4th gen, 1st gen is like banging rocks together. And it is /still/ amazing. You could still, for example, use PCR to amplyfy a particular gene in microarray, with each well being a different sample, and use nanopore electrosequencing to sequence them all at once, very quickly. Mitochondrial genomes are becoming the standard for molecular identification of animal species, for example, and you could sequence a whole bunch of those in no time at all through this method.

Hey, this is the stuff b used to do for a living, before his lab had some funding trouble and he had to go find another job. (The lab did eventually get its funding.)

I don't know the details, I just know that he designed genetic sequencing arrays.

Here's one of his older group papers: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1853129/

[Kai]

the conservative estimates are incredible.  i'm curious what sort of effort would be required to implement them.  is the computation to correlate this volume of data in place already, or would it require additional advancement?
how could this be organized?  is there a medical authority that could handle the task?
it seems the possibilities are a shining jewel.  an irresistible lure.
i'm game.  let's do it!

You would need this generations supercomputers. Despite a human genome being about 750 mb, that's a lot of data points to match. We can cut down on that since we have the human genome mapped, but it still needs a massive amount of processing power. The other issue I suspect will be data privacy. There are going to be problems with piracy and insurance companies screwing people over.

[Kai]

according to the Gizmag article i saw on this, it's a 'disposable' unit.
it also says we can expect the price to drop significantly as production gins up.

Yeah. The DamION seems to be multi use though. Still probably much smaller and cheaper than any pirosequencer out there.

For those of you who are not familiar, there are 3 (now 4) generations of sequencing technology.

The first is Sanger Sequencing, which uses a process called a Polymerase Chain Reaction (or PCR for short). PCR was a radical discovery (apparently discovered after an acid trip) by Kary Mullis.

The general idea is this: you have one DNA strand and you want a whole bunch of copies. Now, you know that when you heat DNA, the double helix pulls apart into the two complimentary strands, and you know that if you added a DNA polymerase (a protein that finds single strands and builds the compliment to them) You will get two double stranded DNA helixes. Now, the problem is most DNA polymerase doesn't like getting heated, it tends to denature. So Mullis looked for bacteria in thermal hot springs and used the DNA polymerase from those. Suddenly, you could add this "taq polymerase" protein which doesn't denature under high heat to the mix, add a primer which will attach to your gene of interest, run the mixture through a successive hot-warm-hot sequence of water baths, and come out with a huge amount of DNA. Every time it goes in the hot water bath, it denatures, every time it goes in the warm water bath, the taq polymerase makes a complimentary strand.

So, now you have a whole lot of DNA. But the DNA isn't the entire length of the gene, because you've added base pairs to the solution that are a little broken, and these will randomly be used to cap the length. Which means you have a whole bunch of different lengths of DNA. And if you include only one type of broken base pair (say, a G (guanine)) in the mix, then all the lengths will be all capped at places where a Guanine would attach. Do this with the other three basepairs, and now you can place each basepair cap in it's own row on a gel electrophoresis setup, and the electric dipoles will pull the shorter strands faster than the longer ones. Basically, you'll have a visual matrix of the sequence, with each basepair in the position by length of the strand. This was updated from the gel setup into capillary tubes, but it still is rather low tech, and requires a huge amount of space to do a goodly amount of sequence. It takes years to sequence a human genome this way. Incidently, this is the method that was used for the Human Genome Project.


Second generation sequencing is somewhat the same, but much faster. It still uses PCR to amplify the genes, and still uses these broken basepairs, but instead of the kind previously, it uses ones that have a dye attached. This is called Dye Terminator Sequencing. Wash away the other basepairs, and you can clearly see the color. Now use an enzyme to cut off the dye terminator, and add another dye labeled basepair. Rinse (literally), repeat, and by the sequence of colors, you will have the sequence of the DNA. But the major problem with this is that, while it's faster than Sanger sequencing, it still takes a long time. You can do this in high thoroughput microarrays, but it still is time consuming. It also can't sequence long segments of DNA.

Fourth generation includes stuff like the nanopore technology. It's fast, it's cheap, and it can sequence very long strands of DNA without cutting them up first. PCR and shotgun sequencing no longer needed to map an entire genome. I hope now that I've described the above methods, you'll understand why my mind is blown. Compared to 4th gen, 1st gen is like banging rocks together. And it is /still/ amazing. You could still, for example, use PCR to amplyfy a particular gene in microarray, with each well being a different sample, and use nanopore electrosequencing to sequence them all at once, very quickly. Mitochondrial genomes are becoming the standard for molecular identification of animal species, for example, and you could sequence a whole bunch of those in no time at all through this method.

Hey, this is the stuff b used to do for a living, before his lab had some funding trouble and he had to go find another job. (The lab did eventually get its funding.)

I don't know the details, I just know that he designed genetic sequencing arrays.

Here's one of his older group papers: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1853129/

That's using hybridization to find single nucleotide polymorphisms (SNPs). This won't even be necessary with whole genome sequencing, because you'll just use software to find the genes of interest after sequencing.

[The Good Reverend Roger]

It's good to see we occasionally get the COOL kind of the future, instead of just the EEEEEEEEEEeeee part of it.

[Scribbly]

the conservative estimates are incredible.  i'm curious what sort of effort would be required to implement them.  is the computation to correlate this volume of data in place already, or would it require additional advancement?
how could this be organized?  is there a medical authority that could handle the task?
it seems the possibilities are a shining jewel.  an irresistible lure.
i'm game.  let's do it!

You would need this generations supercomputers. Despite a human genome being about 750 mb, that's a lot of data points to match. We can cut down on that since we have the human genome mapped, but it still needs a massive amount of processing power. The other issue I suspect will be data privacy. There are going to be problems with piracy and insurance companies screwing people over.

One of the projects my dad worked on was to link up computers through their local area network in order to handle tasks that the company would normally throw at their supercomputers. Overnight, the computers in the office would be left on and the idle CPUs would each handle a small part of large scale data mapping, with some of the machines watching over the others to make sure it all synched up at the end.

I wonder if something similar could be done via the internet. You'd still need a large amount of capital (and knowledge!) to get the project off the ground, but you might not need a multi-million dollar piece of hardware...

Edit: Derp, he was replacing mainframes not supercomputers. My mistake.

[Elder Iptuous]

Sure. that's the idea behind seti@home and folding@home...

[Rococo Modem Basilisk]

Amusingly enough, despite the fact that this device is much faster than anything else currently available, it still doesn't come anywhere near how fast people who watch CSI think DNA can be compared.

[El Sjaako]

Besides helping us with human DNA, wouldn't this help a lot with other diseases? It could perhaps make identifying certain diseases very easy, and then allow us much more accurate data into how viruses, bacteria and fungal infection spread.

I don't know how easy it is to isolate these things from the human, so maybe this is completely impractical.