10 replies to this topic

[Curtis]

Hello everyone,

so here goes my first post after a long time.

 

I work on small RNA viruses and recently applied for a training in Canada to improve my NGS skills. I mean I already had a training back in 2011, and received a certificate for it, but since I did not use the technique I forgot the details and only remember the basics.

 

So I was in the middle of playing with post-NGS software that a friend of mine in the US published an article on sequencing avian viruses using third-generation sequencing. So I was like maybe it is already too late to learn NGS. A newer technique has emerged so no need to go all the way to Canada anymore

 

Most of you may know this sequencing technique as Nanopore sequencing. I remember reading about it back in 2014 that an Oxford-based company was developing a flash-memory-sized sequencing system. I did not pay attention to it much because we had a MiSeq in our lab and I was like this little device could be a hoax. It turns out I was wrong.

 

Has anybody on this forum used this technique? How versatile is it compared to NGS? Do you think it is going to replace NGS in the near future? I have heard its error-rate is higher than NGS.

 

Thanks

[bob1]

I've used it (once) and it worked OK, I am planning on using it a lot more in the near future. The error rate is pretty high (~15-20%) and the software is still in development, particularly for some applications. If you are just wanting to do basic sequencing of samples, then it works pretty well and is reasonably quick (<24h for a run, depending on the read length you want). There are a bunch of caveats to that, in that the technology and software is not quite mature and sometimes needs some user intervention and re-setting at various points. I am of the impression that there is constant development of the nanopore technology and sequencing kits, so the kit you get today might be different to one you get in a few months time. I think this is the only technology that will directly sequence RNA, so that is an advantage.

 

The usage cost is moderately high, as you only get about 10 runs out of each flow-cell, and they are quite easy to damage with air bubbles. I would say that the cost is probably not as high as MiSeq, but you get less information out each run. Having said that, the sequence reads are longer and assembly is easy, especially if you have a reference sequence. For many applications you need a bit of command-line experience, but if you are familiar with those, that shouldn't be a problem (and it's fairly easy to learn). 

 

I do not think this particular technology will supersede next-gen any time soon, but it is certainly growing in popularity and has a lot of followers and a big community support. The company is responsive to problems and suggestions, which is a good sign. Depending on your application you might also want to look at the (much) more expensive PacBio, which seems to be more mature.

[Curtis]

Bob1, thanks as always. I think that RNA sequencing will save so much time and money. I think that is the biggest advantage. It will really be helpful to me. I mean in my lab I still extract RNA, synthesize cDNA, run PCR, clone to pJet and send for sequencing. It is a long process not to mention the time I spend to prepare competent cells and design primers for sequencing the middle regions of the insert.

  

I have read about PacBio too. I mean I saw the name for the first time when I was submitting a sequence to GenBank last year. And then I googled it.

 

I missed the forum, I learned something today. Thanks again bob1

[Trof]

Is it really called "third generation"? ( mind that the "next generation" is already a stupid name, if you consider any further advancement of the methods it becomes weird)

 

There should be something like a table comparing the standard NGS (Illumina, ...), PacBio (longer reads, higher error rate, both of which are used in combination for de novo genome assembly, and this Minions and so.

Also I had personal experientce with a BioNano Irys platform, that uses extra long reads of high-molecular DNA (but not based od sequence, but rather a markers) up to milion bases. Does not give a full sequence thought, but can be used in a hybrid assembly with Illumina or PacBio or both. Currently it is used by NHS to improve the human reference genome, also very popular in plant genomics, they used it to read the complete genome of wheat, plants have often complicated genomes and Irys can read very nicely even through repeats (and through its markers it can define a region as a repeat).

 

So, there would be so many generations of something, that it becomes confusing (never heard of Oxford nanopore as a third though, more like how cool it is to sequence genomes on a something like a flash drive

[bob1]

Yeah, they are being called "3rd generation", as "next gen" was sometimes called "2nd generation"; it is a stupid naming convention, but that's how these things go...

 

There isn't a table for comparison on the wikipedia page, which is where you might find that sort of thing, but the page is pretty basic and not very well written, so I am sure it will turn up eventually.

 

I think the direction these companies are heading is whole transcriptome sequencing, and providing single molecule reads, so that you can infer information directly about each molecule. I don't think the technology is yet at a point where it is absolutely perfect, but there is a lot of active development, and they will get better and more useful as time goes by.

 

It will be very interesting to see what happens in the next few years as these technologies mature and more options become available.

[Curtis]

Who makes these technologies? They are very complicated. The more I learn, the more I realize I know nothing.

 

By the way, they used to call the flashdrive technology next-next-next generation.

 

But what about the software they use to analyze data? Are they shared between different platforms? I was just learning their names that I gave up after reading about the 3rd generation.

[Curtis]

Guys, do you know if the same software can be used for both technologies? what software do you usually use? and do you prefer free online or paid software?

[bob1]

The main differences between the generations is that the data coming out is different lengths - 3rd gen stuff is typically longer read length. I have used the minion, and it comes with some proprietary basecalling software (online, downloadable once you have bought one of their devices, though early versions are on Github). This software can produce files that are in fastq format - which are the standard for Illumina sequencing (others too I guess), so most software works just fine in my (limited) experience. I can't speak much as to commercial usability, though fastq is the recognized standard for this data, so I would guess most will work. 

 

Personally I prefer freeware, as many of the basic sequencing analysis tools are. However, these are often command-line only, so if you are not interested in learning how to do that, you may be out of luck for most of the options 

 

If you are interested in free GUI options and running Linux or on a Mac - try Galaxy, or if you are on Windows try Ugene (cross-platform too).

[Curtis]

Thank you very much bob1. I asked this because I want to be ready before my training starts in November. I want to have software installed on my laptop so that I don't waste time searching and downloading.

[Stephbaker]

I am familiar with the concept of NGS and I have interned with a company so that I could understand the working of the NGS software as I want to work in this field. But I really cannot wrap my head around the concept of third generation sequencing. To be honest I do not understand biology so much, I am from a technical background. Hence, it is tough to understand the newer technologies. I tried looking up the difference between these two and the only useful thing that popped up was wikipedia. Can y’all please give me any links that can give me a clear comparison between both these techniques. The only thing that I am getting from wikipedia is that 2nd gen doesn’t give your longer reads compared to 3rd gen as it breaks down the DNA and decrease the read quality. Your help is very much appreciated.

[bob1]

You have basically understood the difference - Illumina (2nd generation) and the like provide many many short reads - the maximum they can do is about 300 bp, after this the error rate increases. They are basically massively parallel Sanger sequencing - each base added on sens out a fluorescent signal that is interpreted by the machine based on the color of the signal. There is much more complexity to it than that, but it's an analogy that sort of works. 

 

3rd gen sequencing is based off a two different technologies, which can be divided into pore based (Oxford Nanopore) and zero-mode waveguide (PacBio). The pore based technologies work around changes in electrical current with the pore expanding and shrinking as each base is fed through the pore by a molecular motor. PacBio uses a tiny well called a zero-mode waveguide (ZMW) with a polymerase attached at the bottom.  The ZMW is a small area that is constrained in size such that you can observe a single nucleotide at a time in the volume. To do this the polymerase attaches a fluorescent tagged nucleotide to the strand of DNA being synthesized and this is read by fluorescence, as the next base gets attached, the previous one gets shunted out of the ZMW and is no longer observable.