Molecular Technology, From qPCR to 3D Printing
At Festival of Genomics Boston we sat down with Dr Cem Albayrak of Koç University to discuss a particularly exciting technology called droplet digital PCR. Cem was presenting at the Festival on his work on quantifying proteins and RNA from single mammalian cells, an exciting field with enormous potential in the world of both research and clinical genomics. And we even found time to cover another exciting technique gaining traction in the world of chemical biology: 3D printing…
You’re originally a chemical engineer by training, and have found your way into the world of genomics and droplet digital PCR. Could you tell us a bit more about your work in this field?
I’m a chemical engineer and biologist by training, and my main area of expertise is protein expression and protein folding. I came to genomics later on through my postdoc at ETH Zurich, where I focused on single cell quantification of proteins and RNA. There are a number of techniques available to do this, but what we strive to do is correlate the readout of these multiple techniques to an absolute number, and in a sensitive fashion.
The assay that we adopted, called the proximity ligation assay, uses qPCR as the terminal step and hence the threshold cycle number as its readout, but this is a difficult metric to interpret as it measures all changes. So presumably the coefficient of variation that you can measure is 50%, not ideal if you want to look at cell heterogeneity in an isogenic cell population. We wanted to get a more sensitive metric, so we turned to droplet digital PCR instead of qPCR.
So that’s where the assay development stems from, and ultimately what we were able to do was take a few hundred cells and for each of them measure the absolute amount of RNA and protein present. This formed the basis of work that I presented at Festival of Genomics Boston last year.
The development of droplet digital PCR (ddPCR) has clearly allowed you to expand your work into new areas and pursue new avenues of investigation. In the wider community there is a lot of excitement about the potential for ddPCR in clinical applications. What do you see as the future for this technology?
Specifically for droplet digital PCR one of the immediate challenges that I see is the ability to multiplex. It’s like the Daft Punk song, better faster stronger! We want to measure more targets simultaneously and ideally monitor the targets in real time. What I do is endpoint measuring, so we take a snapshot if you will. But what you want to do is monitor a culture or patient samples over time to observe biological phenomena.
What I think will become very useful in clinical application is being able to match different cell populations to disease progression and prognosis. If we have that kind of a map that would be huge for being able to diagnose disease and develop personalised medicine.
For example you could take a tumour sample that you can run through a multi-dimensional gene expression profile, and from that you can determine disease progression or isolate a population that might indicate a patient is more or less susceptible to faster disease progression.
One challenge will be to cut the costs. There is huge pressure, in both the US and Europe now, to drive down healthcare costs, and now we’re talking about personalised diagnostics, which may further increase the costs.
So do you feel that the costs associated with the technology are a critical pinch point?
Definitely if this technology is to progress to the clinic. As researchers we don’t necessarily address this immediate cost because we’re interested in the technical aspects first and foremost, because you’ve got to get the assay working first no matter what the cost. But down the line the cost will have to be reduced.
High throughput will certainly come into play; what I presented in Boston was relatively low throughput compared to other techniques out there. Multiplexing will also reduce costs, all of these enhancements will reduce costs.
Droplet digital PCR has really draw attention as a tool for monitoring cancer progression, as a tool for real-time monitoring of disease. Beyond cancer, and perhaps even beyond disease, what other areas do you see this technology being applied to?
That is a great question! Outside of the clinical applications, immediately I go back to chemical engineering and bioprocessing, and could see droplet digital PCR being applied to monitoring the health of your batch culture, for example. Right now we rely on online or off-gas analyses for certain molecules. But if we know that if you are expressing a certain product then the health of the culture is very important.
What are some of the technologies available at the moment that have really caught your attention, and that could have immediate or exciting implications?
One technology that has really struck me is 3D printing! I think we’re barely scratching the surface. In microfluidics, some groups have attempted to make modular parts that will allow you to put microfluidic devices together like Legos. Right now you need dedicated clean spaces, apparatus, skilled people and so forth to be able to make really large scale complex devices, but I hope that with 3D printing this will change. In healthcare, a pharma company has obtained FDA approval for a 3D-printed pill that achieves a unique release profile for the active drug.
Again, coming from a bioprocessing background, I can imagine being able to print sterile bioreactors straight off your printer, which would be very exciting.
What are some of the major challenges in both improving the digital PCR technology, and in broadening the wider application of the technology?
I’ve touched on this already, but one of the main ones is being able to multiplex. The droplet digital PCR system is currently limited by the number of colours it can analyse. Furthermore, any antibody-dependent technique is limited by affinity, so if you want to multiplex you have to make sure that your antibodies not only have high affinity, but also are compatible with each other. With multiplexing this will become an even greater challenge because the number of unspecific interactions that you have to validate the absence of increases.