Major challenges of nanopore sequencing

[Jeremy]

http://www.nanoporetech.com/technology/publications

There is NOTHING there newer than 2010, which is ancient in that field, and all technology papers are generic reviews circa 2007-2008. Exactly which papers on ONT's website explain how they deal with 256 levels in a 20-30 pA range? Or how they form and maintain the lipid bilayer in a massively parallel fashion? And how they plan on scaling it up? And how they keep the nanopore busy to achieve that spectacular AVERAGE read rate they claim? And how exactly does one deal with temperature and vibration when using a USB stick? And whatever happened to the minion (aka the most unfortunately chosen commercial name of all time)? It seems that they let it eat whatever it wants and it got fat :-)

I guess ONT's absence from AGBT 2013 answers a lot of these questions.

*shrug* what are you asking me for? (I'm not affiliated with them in any way) The papers there did answer many of the OP's questions though. As for your questions, yes, sadly it looks like ON don't have any answers for them.

Shame, it would have been really nice to get sequence data consisting of Mb long reads. They might still surprise us, but it doesn't look like it.

[BBoy]

Some of the papers from oxford nanopore have already addressed many of those problems, rather than just writing that they are a challenge it would be good to write about how those challenges have already been overcome. The oxford nanopore site has a list of publications you could use.

http://www.nanoporetech.com/technology/publications

There is NOTHING there newer than 2010, which is ancient in that field, and all technology papers are generic reviews circa 2007-2008. Exactly which papers on ONT's website explain how they deal with 256 levels in a 20-30 pA range? Or how they form and maintain the lipid bilayer in a massively parallel fashion? And how they plan on scaling it up? And how they keep the nanopore busy to achieve that spectacular AVERAGE read rate they claim? And how exactly does one deal with temperature and vibration when using a USB stick? And whatever happened to the minion (aka the most unfortunately chosen commercial name of all time)? It seems that they let it eat whatever it wants and it got fat :-)



I guess ONT's absence from AGBT 2013 answers a lot of these questions.

[andreask]

I think Oxford Nanopore pretty much proved at ASHG2012 that they don't have any technology that is ready for prime time. Of course, they will claim that they showed the real deal at the conference. But they didn't use it. And come on, why would someone haul expensive, fully working but top secret equipment half way around the world and then not use it? Especially when they claim that their machines are already competitive in certain settings?

[seqnextgen]

Several things that one needs to think about nanopore sequencing even at a layman's level

(1) whether eventually it is possible to have single base resolution in the so-called "strand sequencing" mode

(2) if not, the bigger the foot print of a DNA inside the pore, the many more states one needs to distinguish. There are two related issues, dynamic range of the current measurement and the noise level. One needs to increase the range of the current measurement and reduce the noise to call a definite state. When I was a kid, I was always puzzled by why modern day digital computers do not use more then two states. I think there is a good reason about that. (There are some attractive ideas about new analog/clockless computer, but it is certainly harder to do engineering work.)

(3) Now, if one can not get definite states. Let's say some states are close to a degenerate level. It is in principle possible to use advanced algorithms to improve the final base calls. Have such algorithms already be known well enough such that investing to put them on ASIC/FPGA is worthy already? If not, one will better use a general purpose CPU for such processes. In an academic lab, this is a non-issue. For a commercial product, one needs to consider to the right development paths.

(4) Does a DNA molecule really go through the pore as smooth as shown in the videos/cartoons? Double strand DNA is stiffer but single strand DNA is very flexible. For those who use land-line phone before, one must notice how easy the telephone hand-set coils get tangled. How does such molecule level entanglement affect the ability for reliable signal/data processing pipeline. For example, if one sees the current is stuck at one level for a long time. What is happening? Nothing in the pore? Something in the pore but the DNA is not moving due to entanglement + current level calibration is off? Some modified DNA bases to get itself stuck? The pore is broken? In a cartoon, one does not have to think about all of these issues. In reality, one needs to consider all possibilities.

(5) Is nanopore sequencing robust against any impurities in the sample? Do salt and other unwanted molecules affect the sequencing results? It is hard to believe one can take blood and get high QV non-random DNA sequencing data without any sample preparation. It seems deifying the 2nd law of thermodynamics.

All of these are really interesting problems. Maybe nanopore seq. does not have all these potential issues, however, only some real data can prove it.

By the way, although I have some knowledge about general statistical physics, I have never worked on any nanopore exp/data analysis. I can be totally wrong about nanopore sequencing and I am looking forward to the technological advancement to solve all potential problems. It is indeed an interesting time to live to see how to reveal the mysteries about life piece by piece.

[cibertech]

Hi nanoporous

Thanks for your message.
I am aware of those groups. It aseems the talk from prof. Slaven Garaj this morning, at NGSC about it was also interesting.
But all info on this is welcome.
I will send you a PM so we can talk.

Thanks and best regards

Ezequiel

[Nanoporous]

Also check out the recent publication by the Gundlach group on addressing the speed issue using a phi29 polymerase.

There are also other approaches being taken by other companies for reducing the electrical noise, and slowing DNA motion that does not involve enzymes. Send me a PM me if you want to talk about it.

[cibertech]

Thanks

Hello Jeremy

Thanks for your comment.
I do agree with you. That’s actually what I am doing right now. As I said, I’m writing an article about this and I have to start somewhere, so I have started with the problems, and will follow with the current solutions, not only from Oxford nanopore but also from other research groups currently working in this area.
I just want to make sure I'm not forgetting any issues and also don’t have any errors. With almost 150 visits and only your comment, I believe I must have done a very good job!

Thanks again

Ezequiel

[Jeremy]

Some of the papers from oxford nanopore have already addressed many of those problems, rather than just writing that they are a challenge it would be good to write about how those challenges have already been overcome. The oxford nanopore site has a list of publications you could use.

[cibertech]

No differentiation made between strand sequencing and exonuclease sequencing methods
Also, no distinction is made between biological and solid state pores

Major challenges of nanopore sequencing.

One of the biggest challenges is electronic noise. The electric current changes in the nanopore are in the range of 25 to 50 pA (25 - 50 x 10-9 A) and that is for the difference between having the nanopore empty or having DNA passing through it. Imagine how difficult it is to differentiate one base from the next!
Also, these are extremely small. Nanopores (natural and synthetic ones) can have diameters between 1.2nm (MspA pore) and 50nm (solid state). This presents a challenge for the interface, as the nanopores need to be connected to an outside circuit that processes the signals. If that wasn’t enough, the lipid bilayers on which biological nanopores are suspended are unstable and hard to manipulate. They are also very fragile and can break easily. As soon as that happens there is no more voltage between either side of the nanopores and thus translocation stops (i.e. the DNA does not move).
There are also uncertainties about whether translocation occurs at a constant speed and the complications of pore clogging.
Another issue is the speed at which the DNA strand moves through the nanopore: at a rate of 1 to 5μs per base. Some companies are trying to address this by controlling the speed of DNA strand by various protein engineering strategies, using phi29, for example.
One specific challenge of the exonuclease approach, where a processing enzyme chops individual bases of the DNA and feeds it to the nanopore is to make sure the released nucleotide goes into the nanopore. This is being addressed by attaching the enzyme to the nanopore, although this mechanism does not constitute a guarantee that the nucleotides will follow a particular path.

[cibertech]

Hello all forum viewers and participants.
I’m trying to put together an article which talks about all the challenges currently faced by companies trying to develop nanopore sequencing technologies. (Please note I am not refering to any specific company here)
Please feel free to comment on my findings below, add or correct if needed.

Thanks and best regards

Ezequiel