Feature: Sequencing 3.0

This feature appeared in the May/June 2011 issue of Australian Life Scientist. To subscribe to the magazine, go here.

It is not often that one can read a company’s blurb about their latest new machine-that-goes-beep and not be at least a little sceptical about its purported capabilities.

But after talking to Professor Sean Grimmond, Director of the Queensland Centre for Medical Genomics at the Institute for Molecular Biosciences, and taking a look at the different sequencing machines he currently has operating, Ion Torrent’s claim of “democratising research” and having a sequencing machine on every lab bench sounds entirely reasonable.

In December 2010, the Ion Torrent Personal Genome Sequencer (PGM) officially went on sale. Labtech company, Life Technologies, bought Ion Torrent late last year and with that came the PGM, a revolutionary (really this time, for a change) step in DNA/RNA sequencing.

For a start, it is about the size of a benchtop centrifuge, although much prettier with some very funky design aspects, and costs around US$50,000. But as Grimmond explains, the size and appearance is by far the least of its advantages and he for one is already a big fan.

Grimmond leads Australia's effort into the molecular analysis of pancreatic and ovarian cancer as part of the International Cancer Genome Consortium (ICGC). He has a long and internationally recognised history in microarray expression profiling and bioinformatics, which coupled with his more recent move into medical genomics makes him very well qualified to comment on the newest kids on the sequencing block. Besides that, his was the first lab in Australia to get their hands, pre-market, on the Ion Torrent technology.

All of the current platforms including the Roche 454, Applied Biosystems SOLiD and Illumina’s Genome Analyzer all detect and determine the base reads using optics. According to Grimmond, the Ion Torrent technology does the same thing just without the optics. So no fluorescent labels, precious lasers or temperature and vibration-sensitive cameras.

Instead, the magic happens via a good ol’ silicon chip. Yes, that’s right – the same semiconductor technology that has been around for 20 years and which drives almost everything Western society cannot live without from computers to phones to our cars – now it also reads your DNA.

Ion Torrent sequencing still uses the conventional chemistry and a bead-based solid phase, whereby the DNA molecules for reading are attached to beads and put into tiny conical wells, and the sequencing is done by synthesis using polymerases. The wells are then flooded with each nucleotide base in turn (C, G etc.), and these are chemically incorporated or not by the sample depending on the sequence present.

---PB---

However, instead of measuring the fluorescence reaction as each base is incorporated, the Ion Torrent measures changes in pH generated by the hydrogen ions that are released during the process of nucleotide incorporation.

“Relative to how much of a particular base is added, you get a quantitative difference in the amount of hydrogen ions released (hence the name),” says Grimmond. “Then, as you systematically put through each nucleotide, the hydrogen releases are recorded as pH spikes, and you end up with the base call that is, the nucleotide sequence.”

This information is then 'written' electronically onto a silicon chip, so the base information is converted directly from chemical into digital data literally in seconds. The actual machine is therefore like a very sophisticated pH meter inside a big box with holes on the side to pour in the nucleotides. The ‘chip’ inside comprises millions of tiny wells for the samples sitting on millions of tiny electrodes – very clever stuff and yet seemingly so simple.

Smaller, faster

According to Grimmond, having no optics and using the direct chemical to digital detection gives the Ion Torrent several pluses. “The data file sizes are small, and the way it actually analyses and measures the nucleotide incorporation is quick, generating reads of about 120 bases takes less than two hours.

And, most importantly, the semiconductor technology means the potential to scale far faster and certainly far further than any of the currently available sequencing technologies.

“With any of the current chemistry approaches, the files you need to collect are really quite huge. Converting changes in pH directly into a base call means much smaller files sizes. You really could run those machines pretty well all year without emptying the hard drives, whereas we run the SOLiD machine twice and then we have to move data to make more room. This is a huge issue for this type of work and it is only going to become bigger.”

The first-release Ion Torrent machines come with what is called a ‘314’ chip, containing about 1.2 million wells and matching electrodes. A newer ‘316’ chip (6-8 million wells) is due out any day, and within a year Ion Torrent is planning to release the ‘318’ chip which, according to Grimmond, will comprise ~25 million wells.

“So, using the exact same machine and sequencing platform, you can go from generating 1 million base reads to 25 million reads, and with that many wells, we are getting into the one giga base pair range of data in around two hours with relatively small file sizes. As a comparison, the SOLiDs right now can generate about 100 Gbase, but that takes about 2 weeks and masses of data space.

---PB---

“So, if I want to sequence a human genome for example, we would still be using the SOLiD, because you need that sort of volumes of data. But for smaller and more tractable sequencing applications such as the transcriptome or microRNAs, or candidate DNA mutation analysis, or microbial genomes, the Ion Torrent PGM is ideal.”

The Ion Torrent machines have already become a very useful part of Grimmond’s sequencing armoury. In the ICGC program, the PGMs are running solidly to validate mutations picked up in patient samples sequenced on the Centre’s ABI SOLiD workstations and to address questions of clinical significance in those patients. The technology is ideally suited for these applications that take up a huge amount of time and space on the genome sequencers, but can be run in a fraction of the time using Ion Torrent.

“For example, we can now do very deep sequencing on samples from the tumour margins using primers that will detect every mutation found in the parent cancer, and in this way more closely define the risk of metastases – this is particularly critical in the case of pancreatic cancer.”

The PGMs are also helping with the massive job of validating every DNA variant found in the cancer genomes. “We can cut some corners to pick up some of those variants, but for the novel ones we really need validate around 200 mutations per individual. Ion Torrent allows us to automate our primers, hone in on the regions that we think will have mutations, PCR them up and sequence them all on the chip and then move on to the next one – quickly and easily,” says Grimmond.

“This technology has really completely changed the dynamic of how we do the next-generation sequencing discovery then validation phase, and it is actually cheaper to do than microarrays were not all that long ago.”

Ironing out the bugs

The new machines are also being put to good use on sequencing jobs outside of the cancer program. “We have started to do a few bug genome sequences, working with Professor Matt Cooper here at the Institute looking at pathogenic bacterial strains.

The fact that the reads are quite long with the PGM, and just getting longer, means that we can avoid some of the challenges you may encounter with some of the large high-throughput technologies, which at best will only give you 100 bases or so out of each run.

---PB---

“But now with the Ion Torrent, we could potentially get 200 up to even 400 bases out of each run. And this completely changes the way we use that sequence data – easier to assemble and to analyse.”

Grimmond’s group are finding that they can now undertake very rapid method development using the Ion Torrent machines that usually would not be feasible because of the sequencing time and fiddling needed for optimisation steps on the SOLiD stations, which run 24/7 on the cancer genomes.

“For instance if we are going to look at some new population of DNA or RNA molecules that someone has managed to isolate through some new funky way in the lab, we can easily test for things you need to know along the process.

Having a machine that will generate 20-100 megabase of sequence easily, cheaply and quickly opens up a lot of opportunities to what and how you might want to sequence that would be neither feasible nor affordable otherwise.”

This ultimately enables entirely new questions being asked and quite different pipelines being established. “I think these machines will become an important tool in the sequencing world.

“It is quite easy to see that if they can make a silicon chip that determines DNA sequences the size they are now and sell it for ~$200, and you can generate enough long reads, it would be very easy to make a bigger chip that could generate a human genome in two hours. The detection system needed is virtually already built – they just have to work out how to get the molecular biology down to fit in with the more and more sophisticated chips.”

Whatever the outcome, next-generation sequencing is certainly a very exciting space to work in, and Grimmond expects that with upgraded conventional systems also coming out in the near future and innovations like Ion Torrent pushing at the big boys’ heels, “we will be reaching data sizes of ‘biblical’ proportions in the near future - then you really might start seeing one on every bench.”

This feature appeared in the May/June 2011 issue of Australian Life Scientist. To subscribe to the magazine, go here.