Small particle sorting in flow cytometry and the need to integrate ‘omics’ data- Interview with Dr. Oscar Fornas, Pompeu Fabra University and the Centre for Genomic Regulation, Barcelona, Spain.
This month, www.flowcytometry-news is proud to present an interview with Dr. Oscar Fornas from Centre for Genomic Regulation and Pompeu Fabra University in Barcelona, Spain. With nearly thirty years of experience in flow cytometry, Dr. Fornas is one of the leading practitioners of small particle flow cytometry in Europe. In this interview, he spoke to us about his research, the challenges involved in small particle flow cytometry, his view of multi-omics and the importance of correlating data obtained from different techniques.
1. Can you tell us about the current research in your lab, what you are working on?
I am the head of the flow cytometry unit, which is a joint one run by two institutes. These are the Pompeu Fabra University (UPF) and the Centre for Genomic Regulation (CRG). I have been running this facility for the past 18 years and we are currently managing about 200 users. More than 120 research groups use our facility. We have users and companies not only from the Barcelona Biomedical Research Park (PRBB), but from all over the world. We provide our experience, expertise and training, and access to six state-of-the-art analysers and two cell-sorters. In research, we are continuously scouting the technical needs of the different researchers and developing new applications for flow cytometry in collaboration with them. Consequently, we are continuously providing technical solutions in flow cytometry. Currently, we are focused on sorting small particle detection and isolation. There are several people interested in sorting small particles like extra-cellular vesicles, viruses and chromosomes. One application we have successfully developed recently is virus isolation for single-virus genomics1. This is a very interesting approach because it enables the isolation of single-viruses for genome sequencing. As a result, we can study the whole virosphere to understand the global microbiome. This is an interesting example of how being in close contact and scouting the needs of researchers from a facility can result in technical solutions for the whole scientific community. Another interesting application we have recently implemented in our facility is chromosome sorting.2 As you know, this is not a new application of flow cytometry and is actually one of the oldest. I was very interested in implementing this one because we now have had approximately two decades of whole-genome sequencing and the huge amount of generated data is still difficult to analyse. I was interested in fragmenting the genome in order to focus on each chromosome in detail and, specifically for complicated chromosomes like the Y-chromosome, flow cytometry is a very interesting tool. About twenty years ago, it wasn’t possible because of the budget, but finally, around ten years back, it became possible to acquire a BD Influx™ with the proper configuration, and we investigated the possibility of studying chromosomes. We implemented this application in collaboration with a researcher from Yale University who was put in touch with us. Finally, now we have several different projects based on chromosome sorting. We recently published a paper where we isolated the Y-chromosome1 and its subsequent sequencing by using new MinION Oxford Nanopore® Technology. This was done in collaboration with an internal researcher who works on evolution. Due to this, we are receiving a lot of collaboration offers. I think we are the only ones that do this in Europe as a service in collaboration for researchers. Now we are going back to the past with this tool. As I mentioned, it is not new, but it is becoming a very useful tool to interpret and analyse the sequencing of the different complex parts of the genome, i.e. Y-chromosome. Through the combination of sorting and Nanopore® technologies, we are directly sequencing the DNA without amplification. This is now becoming a more interesting way than conventional sequencing techniques.
2. What are the specific challenges involved in sorting small particles like chromosomes and viruses?
Chromosomes are not really small but the challenges in sorting them are not only to have the right configuration in your instrument but also the need to have someone who is an expert in managing the cells and preparing the samples. It is crucial to know if you have the samples prepared properly. In our case, after several years now we have developed this expertise. The challenge then is to set up the cell sorter properly. We get a lot of support from BD, which is helping us. They initially put us in contact with different experts in the field, for example in Utrecht in the Netherlands and in the US. This was very useful.
For viruses and exosomes, which are very small- we are talking about nanoparticles, we used cryo-electron microscopy to elucidate the size of the particles that we could sort. We know that we need a resolution of 30-40 nanometers, which is very high. The challenges are great because you never know if you are identifying and sorting these very small particles correctly. We have had to use complementary techniques like molecular biology and microscopy with lots of different controls. All these are very crucial when dealing with small particles.
Something that I can tell you is that surprisingly, contrary to popular opinion, in our hands, the analog systems like BD FACSCalibur™ and BD Influx™ resolve small particles much better than the new digital systems like BD FACSAria™ or BD LSRFortessa™. This is something I’m continuously discussing with BD and for me, it’s a pity that we cannot have something like a BD FACSymphony™ with a possibility of switching between digital and analog, as was with the first digital BD instrument, the FACSvantage DiVa™. This is only a question of how these instruments process the signal.
3. What are your expectations from a company like BD? How can we help you with your research? You’ve already talked about analog sorters but is there anything else we can do?
I started to work with BD around 20-25 years ago and my experience has been really good. I always recognise that their technical service has been amazing and we have been receiving a lot of help with the technology we are managing. This not only concerning software and applications but also on the working of the instruments. For me, this is important because when you are trying to approach the limits of the resolution of the technology like in exosomes or viruses, it is important to understand and control the technology you are managing. All the engineers from BD have been very helpful to us. Thinking about the future, it is good to know that BD is continuously advancing, scouting and developing. I like the fact that BD continuously listens to us, takes our opinions seriously and implements technological advances in the instruments. For me, something that would be very interesting is what I have mentioned before – the possibility to manage different technologies like analog and digital and if they could be interchanged. As an expectation, I would like that they keep developing technologies. Sometimes the technology development goes faster than the real possibilities which is good as you get more tools than you really need. Now BD has FlowJo™, which is a very nice thing to know. As we are going towards high-dimensional flow cytometry and are addressing very complex immunophenotyping, new tools for this are a real need and it is good to know that BD is developing them. We are also now generating tons of data and I’m afraid that we are making the same mistakes that the genomics people did 10-20 years ago. They generated lots of data and are now contracting bioinformaticians as they don’t know what to do with it. We seem to be going in that direction in flow cytometry. High-dimensional flow cytometry data is difficult to analyse and we are facing big problems with our complex data. So it is good that programs like FlowJo™ have become available. Generating data is one thing and getting biological information from it is another. Another thing I am expecting from companies like BD is to expect the unexpected. Post-sales support is essential and help with tools is crucial as there are a lot of people interested in high-dimensional flow cytometry. We expect new highly advanced tools
4. We have recently introduced multi-omics technology. What are your views on this, and would you consider using this technology in your research?
Continuing with our previous discussion on complex data, when we go for other ‘omics’ like genomics, transcriptomics, etc, we need to integrate all the information. The companies developing these technologies need to provide ways to integrate the data. I know about the transcriptomics and genomics that BD is doing, and tools in order to further integrate the single-cell analysis, phenotype and the transcriptome are coming. This is the real future because now we need to integrate all the ‘omics’ in general to have a complete biological picture at the single-cell level. The keys to the success of technologies like multi-omics are in the hands of companies like BD which are developing the tools that we really need. In fact, it is not the future anymore as we are right there.
5. My next question is as to what are the burning questions in single-cell genomics and what direction do you think will the research take in the future?
It is again related to the previous discussion. At the single-cell level, we now have a lot of tools to isolate cells. We have been managing technologies other than flow cytometry, but people seem to be returning to it as it provides the best reliability and accuracy for single-cell isolation. Other technologies are giving us artifacts or at least, we haven’t got the reliability we need. When we go to single-cell genomics at the end, or even transcriptomics or any other technology, we need to co-validate all the results and they should be completely correlated. Evidently, reliability is crucial. Since genomics goes in one direction and transcriptomics in another, phenotyping and cell biology in yet another, the most important thing is how to integrate all of them. We need sophisticated tools that can handle and integrate all these techniques. Developing new tools, in terms of software, is crucial. We have tools to validate flow cytometry, but we need better tools to validate the other ‘omics’ and most importantly, be able to validate the integration of all of these ‘omics’ properly.
6. We have this discussion a lot, as to how we can make multi-omics complement flow cytometry in the kind of data that it gives, as it allows you evaluate simultaneously both mRNA and protein expression while flow gives you just protein marker expression. How multi-omics can complement flow cytometry is a big area we are pushing on.
This is where most of the efforts must be put.
7. What is your opinion about spectral cytometry and do you feel BD can help in this area?
We have a spectral analyser from Sony®. This is a complementary piece of equipment in our lab as all our other instruments are from BD. Along with the BD FACSCalibur™ and the BD LSRFortessa™, we purchased it three years ago, as it was the only existing tool then that could cover high-dimensional flow cytometry. Spectral cytometry is a different approach to phenotyping. The idea of it is ‘spectral’ is not a revolution. The main difference is how we can manage more than 20 parameters at the same time, that is the revolution. I know that the new BD equipment is going in that direction, to high-dimensional flow cytometry. There’s another company Cytek® using a technology that is doing something similar. It collects signals from the whole spectrum but is based on band-pass filters and detectors. Consequently, you also analyse the whole spectrum, in a different manner than the Sony® one, but addressing spectral cytometry. In my opinion, it is not important if it is spectral or not. The important thing is that we can manage a lot of parameters, particularly more than 20 that people are interested in. It is also interesting how they have developed algorithms for the deconvolution of the highly overlapped spectra. Then, now we have the opportunity to handle different fluorophores that we couldn’t do in the past.
The interesting thing about spectral technology is that it has pushed other companies to develop this technology even faster to cover a lot of colours at the same time, so we have more possibilities with more companies. Of course, being able to manage a lot of parameters is not as important as the analysis of the data and the biological insights that one can gain from the experiment. As I have already mentioned, it is very difficult to analyse high-dimensional data and this is a real problem. In the end, you need to know what you can conclude at the biological level, not just the technical level. We are just not able to understand the complexity of phenotyping. When you asked me what I am expecting from a company like BD, for me it would be interesting from a technological standpoint in the coming years if you can provide us a cell sorter that can carry out clustering and t-SNE analysis in real-time. When I have a specific and new strange clustering then, even though I can believe the analysis FlowJo™ is doing, I cannot say anything about that phenotype and its biology if I cannot isolate that specific cell population and confirm it using other techniques. I think it is very important for us to have a cell sorter that can provide real-time clustering and sorting. That will be amazing.
We did the data analysis course with the FlowJo™ people about a month ago and were discussing the t-SNE analysis and other tools. It all looks nice but we must do more. To get the biological relevance we need to isolate the populations. Only then complex phenotyping will be biologically relevant. I know this is difficult, but I also know that companies are working in the direction and this is really great to know.