An interview with Professor Matti Nykter: CSO and dedicated researcher

In this session, we take a deep-dive into the mind of Professor Matti Nykter – co-founder and Chief Scientific Officer of Genevia Technologies, and Professor of Bioinformatics at Tampere University. Together with Prof. Matti Nykter, we gain unique insights into the mission of a bioinformatics company, the most promising trends in computational biology, and valuable tips on how to make the most out of cross-disciplinary research.

Venla Kiminki: First of all, thank you for making time for this interview, Matti! Let's get straight to the topic of the day – you. You began your career as a computer scientist, then turned your hand to computational biology, and eventually became a professor of bioinformatics. Can you give us a bit more background into this fascinating academic career trajectory?

Matti Nykter: “I became interested in computers and natural sciences during my early school years and studying computer science at a technical university felt like a natural choice for me. Having studied signal processing, mathematics and information technology, biosciences were not an obvious career path for me. However, I began to notice how molecular biology research was becoming more and more a data science with the advent of microarray platforms, and the researchers there needed to apply methods I was very familiar with like statistical analysis and mathematical modelling, not to mention scientific computing technologies to do all this in general.

After graduation with a Master’s degree from computer science, I joined the computational systems biology research group at Tampere University of Technology to apply my skills in biomedicine. During my graduate studies, I got an opportunity to work with bioscientists who got me interested in applying my skillset in biological data analysis, especially on cancer research. After my post-doc in the US, I returned to Tampere University of Technology as a group leader and was subsequently appointed as a professor of bioinformatics at Tampere University, Finland.

Today, besides being the Chief Scientific Officer of Genevia Technologies, I lead a cross-disciplinary computational biology research group. My research group employs around 30 scientists – from molecular biologists to computer scientists and mathematicians. The group is part of the Academy of Finland Centre of Excellence in Tumor Genetics; we study how DNA alterations and epigenetic modifications bring about transcriptional changes that result in tumorigenesis and the hallmarks of cancer progression from the interplay with the immune system to metastatic spread. As a bioinformatics research group, we are to some extent agnostic when it comes to different types of cancer, but often our model of choice is prostate cancer.”

It sounds like you enjoy applying computer science and data analytics to biology. What aspect of bioinformatics most excites you these days?

“Combining high-powered computing, high-throughput experiments, and bioinformatics algorithms to shed light into biology that is hard to study with traditional methods. I guess due to my background, I am fascinated by trying my hands at all the various biological questions to which only bioinformatics can provide the answers. For the past years, I have been preoccupied with deciphering the multi-level networks and processes that make cancer cells tick. For example, I am extremely interested in analyzing the huge amount of data we can get from cancer patients to learn how the individual’s disease develops. This is something I hope can be translated into personalized healthcare solutions by medical researchers. I can’t wait to see what else we find in the next few years when biology and computational sciences become even more tightly knit together.”

From your uniquely qualified vantage point as a professor in the field, what do you see as being the most prominent bioinformatics trends in 2020?

“Developments in Bioinformatics are often driven by technology. From a sequencing technology point of view, platforms enabling the sequencing of longer reads have been in development for quite some time, but now these technologies are finally starting to break out. This year the first end to end assembly of a human chromosome was published. Thus, we are moving towards more comprehensive identification of e.g. mobile and repetitive elements in personal genomes. Alongside single-cell technologies breaking out and scaling towards multiomics from single cells, the sensitivity of sequencing is increasing, for example, by leveraging innovations like unique molecular identifiers. Thus, more broad coverage and high sensitivity enable new applications in both basic research and diagnostics, both of which present bioinformatics challenges that need to be addressed to enable translation.”

Given the many promising scientific developments that are being worked on at the moment, what breakthrough advances do you expect to see in the next five years?

“Computational biology is taking huge steps ahead all the time, so it’s extremely difficult to predict where the field will be in five years. Present-day technologies, like single-cell sequencing, will most likely advance to enable probing cellular behaviour in their natural environment. In cancer research particularly, that would allow us to have a more holistic view of the entirety of the biological systems created by these cells – the tumors. From the computational biology point of view, that also means developing new models that take spatiality into account. When high-throughput molecular experiments are taken into routine clinical use, I expect advances in personalized medicine improving the patient’s prognosis as well. This would mean computational biology becoming much more mainstream and practical than ever before.

The basis for these advances has already been prepared. For example, the International Cancer Genome Consortium ICGC has already mapped out the common factors for cancer development, and we already have a pretty good idea what type of alterations initiate cancer. The next step is to better understand what makes cancer fatal. Prostate cancer is a good case-in-point for studying this question: most men develop prostate cancer when they age, but why do some suffer from a lethal type of the disease while for most the disease doesn’t affect their everyday life at all? Developing a more sophisticated understanding of cancer biomarkers indicative of a more severe clinical course would also save a lot of medical expenses as the treatment would become personalized.”

You co-founded Genevia Technologies back in 2011 when sequencing was still prohibitively expensive for most researchers; what was your motivation for introducing a bioinformatics service back in those days?

“Back when I started my career in bioinformatics, high-throughput experimental platforms such as microarrays, next-generation sequencing, and parallelized mass spectrometry data were scarcely available, expensive and in general not as popular as they are today. As microarrays became cheaper, more reliable, and easier to outsource, soon enough every other molecular biology research group seemed to have their own data set. During those days, bioinformatics was kind of a niche area of research, but the advent of microarrays suddenly skyrocketed the demand for biological data analysis skills. My research group started receiving a lot of collaboration requests from bioscientists of all kinds, and soon we found ourselves constantly pursuing data analysis projects.

After years of providing data analysis support to bioscientists through collaborations with my group, we started thinking that there must be far more bioscientists in need out there than we could ever collaborate with. We thought that researchers with more straightforward data analysis needs – who were the majority, actually – would benefit more from dedicated bioinformatics support than academic collaboration with me and my students. Serendipitously, I learned that there was an academic funding instrument available for looking into the commercialization of research which we decided to apply, and subsequently got awarded with. With the help of this grant, we took some time to seriously investigate the commercial potential of providing bioinformatics services to bioscientists in academia and industry.

As a result of this project, we decided to found a company with my first postgraduate student, Antti Ylipää, to provide essentially the same service to essentially the same people that we were used to working within an academic setting. Genevia Technologies was founded in 2011 as a sort of a commercial bioinformatics core lab to help bioscience research groups with their microarray data. As this was still in the early days of high-throughput experimentation in biosciences, the business started inching forward, but it was a rocky road for academics-turned-entrepreneurs. Bioinformatics services were a lot harder to sell than we had expected since no-one was used to buying data analysis, had budgeted any money to cover data analysis costs, or even understood the benefits very well. The go-to option for most academics in the early 2010s was still to look for an academic collaborator who would be willing to analyze their data in exchange for authorship in their paper, like we had done.

The main catalyst for Genevia Technologies was the next-generation sequencing revolution. During the golden age of microarrays, next-gen sequencing – our current focus – was still prohibitively expensive and practically out of reach for most researchers. However, as suddenly as the microarrays had broken into the mainstream of molecular biology, crashing sequencing prices and increased availability resulted in NGS data taking over the whole field by storm. Bioscientists were suddenly drowning in a wealth of ever-increasing amounts of data. Luckily for us, this also meant that researchers had got used to the idea of outsourcing lab experiments like sample preparation, sequencing library construction and the sequencing itself; sometimes they even outsourced the bioinformatics as the option became available from the sequencing service providers.

The social stigma of outsourcing a part of one’s research cycle started to wane as academicians started thinking about industry collaborations as more of an everyday tool to increase the impact of their scientific work, rather than something to be embarrassed by. Bit by bit we were able to get our foot in the door, and few forward-thinking scientists started to appreciate working with NGS data analysis professionals rather than students. Only then did we really start to appreciate the fact that having spent time in the leading academic institutions in the US, and collaborating with experienced bioscientists over the years had given us world-class expertise that still acts as the solid scientific foundation of Genevia Technologies.”

What impact have 10 years as a bioinformatics entrepreneur had on how you approach your academic work? Have you acquired any insights from the industry side that you have found to be valuable in your research?

“Well, I’ve learned that you can leverage the incredible scientific knowhow that trickles from academia into companies in the form of postdocs taking industry jobs. Many of my academic colleagues are disappointed when their most seasoned students leave academia, but that does not mean that the best and brightest cannot keep contributing to your research. Instead of paying them a salary, you just have to start paying for their services to a company or pay for the research products they design and manufacture. Now that I have a much better grasp of what industry can offer, and how to work with companies efficiently, in my academic research we aim to outsource everything that can reasonably be outsourced.

For example, even though our university has a – rather small – next generation sequencing core facility, we outsource basically all of our sequencing to companies. This is mostly because we want to spend all possible resources in the work that makes my group efficient - and keeping up with investing into the latest instruments and sequencing protocols is certainly not one of those tasks! My group is a part of a Centre of Excellence, and thus we are relatively well funded, but I still prefer to use companies if they can bring me added quality, speed and cost-savings compared to a more traditional DIY research. Like everyone, I naturally let companies make the reagents, pipettes, and devices I need for molecular biology experiments such as preparing tumour samples for single-cell characterization.”

Moving forward, how do you think the recent advances in sequencing technologies and other omics platforms will affect bioinformatics in general, and more specifically the scientific methods harnessed by Genevia Technologies?

“I predict that high-throughput experimentation will become even more affordable and more easily available. That fact combined with advances in algorithms and computation will lead our projects to become more integrative and more complicated. In practice, I see our clients producing more data from more samples - and especially more layers of data, such as sequencing the genome, transcriptome, and epigenome of their samples. This will result in a growing need for those who can analyze and interpret the complex data. In many fields, the low-hanging fruits are becoming more and more scarce, so the trend is moving from simple mRNA-seq experiments to creating data in a more systemized manner. Whereas earlier we might have had a project for profiling a few tens of primary tumors, in the future, we will have projects where we take public data from the Cancer Genome Atlas, for example, and conduct the same investigation with an order of magnitude more data. Hence, the projects will probably extend to longer-lasting collaborations rather than short one-off analyses that will not allow our clients to publish in high-impact journals as easily as before. I hope this will also act as a wake-up call for those groups that have not budgeted funds for the data analysis phase of their research before.”

There seems to be a theme running through your academic research, as well as the industrial side of your work: cross-disciplinarity. What is your recipe for a successful cross-disciplinary approach to life science research?

“First of all, the bioinformaticians’ technical skills must match the level of skills that other collaborators have in their respective fields. The bioinformatics partner must have a good command of the available tools and methods and understanding of how to use them in different applications. Often, I see experimental group leaders setting up research teams where all the experimental scientists are superb, but the bioinformatics is done by a student or an experimentalist who has learned the basics of R. This to me makes about as much sense as to have bioinformaticians running their experimental labs after learning the basics of pipetting. No one in their right mind would even think about doing that, yet the converse is all too often true due to lack of understanding and poor budgeting in grants.

When the technical knowledge of the team is on an equal footing, almost anything can be done. My second point is the level of cross-disciplinary understanding between the team members. If the bioinformatics partner does not understand the biological application, they cannot effectively collaborate with experimentalists in building science on top of the existing research. Unsurprisingly, this also applies to the experimentalists: if they do not understand how bioinformatics has taken biological research forward in general, they will struggle to use bioinformatics as an effective tool in their field. Both sides need a certain open mindset to allow for constantly interpreting and analyzing what was done and if the results come out as expected, like any scientist should.

Third, I want to emphasize the importance of research management. All too often, research leaders give free rein to individual members of the team instead of steering the ship to the port of destination. This should be done to avoid the situation where two groups are working on two different projects apart from each other, even though their end-goal should be the same paper. It is fine to make two or more different papers out of the same subject, but that isn’t inter-disciplinary research then, is it? Hence, fluent communication and respect for each other’s competence are key. It is not sensible to expect that a bioinformatician who has spent 10 years studying bioinformatics has somehow also become a medical doctor, and similarly it is not sensible to expect that a medical doctor can do with R what a trained bioinformatician could do. In the best kind of collaboration both sides, biologists and bioinformaticians, have respect and an open mind to learn from one another.”

Finally, what advice would you offer your fellow scientists on effectively integrating bioinformatics and other cross-disciplinary collaborations into their research?

“Embrace the fact that in the 2020s, you don’t need to do everything by yourself. In fact, if you are still in the do-it-yourself mode, you may find other groups overtaking you left and right. For example, with my research group, everything we do is built on top of someone else’s work; most of the algorithms and analysis tools were made by someone else, and there is only an occasional need to create analysis software by ourselves any more. The same principle applies to experimentalists - you probably don’t have a real need to mix your own reagents, build your own experimental devices, and process terabytes of omics data on a cloud computing server. Someone else can do all these things for you. Even if you personally enjoy spending time doing the basic steps yourself, aiming for high scientific impact and keeping the grant funding flowing these days requires effective collaboration between disciplines, and between academia and industry, especially because your competition is already on the highway of science.”