What is your professional background?
I come from the lab: I studied molecular biotechnology in Munich. During my master’s I became interested in bioinformatics, data science and software engineering. In the beginning of my PhD I tried to do experiments in the lab, working with non-coding RNAs in embryonic stem cell development. We tried to push cells in a defined direction to get insulin producing cells as a therapy for diabetes. I created a few nice videos of differentiating cells but I wouldn’t say that my lab work produced menaningful results. I always had the goal to use data from my own experiments for data analysis and modeling so I accepted ‘lab defeat’ and went full bioinformatics. I continued working on data integration and ‘network biology’ in the area of non-coding RNAs.
I started working with graph databases in 2013 and after the 2.0 release of Neo4j I built the first graph database applications.
What are you currently doing, what experience do you bring along?
With my company Kaiser & Preusse I build custom graph database applications for clients in pharma, healthcare and research. The second area I focus on is data harmonization and data standards (things like OMOP CDM and i2b2). Both topics are connected: I think that Neo4j is ideally suited as a central data hub that connects all the disparate data sources in an organization. Data harmonization is an important step to make unstructured data accessible before you can integrate it with other data sets.
Do you have a special hobby/passion - what do you like to do most in your free time?
I lived in Munich for almost 15 years, so I like hiking and all other kinds of activities in the mountains of course. I recently moved to the north of Germany, to the beautiful town of Stralsund. It’s located at the cost of the baltic sea and my goal for next year is to learn sailing. Considering climate change, it might be a more sustainable hobby than skiing. I also play golf but that collides with what I like to do most: Spend time with my family and my son Emil. We build a lot of Brio railway tracks at the moment and he loves cycling.
My other not so secret passion is politics: I worked as district chairman for my favorite party, organized political campaigns and tried to get breeze of fresh air into the sometimes arduous party life. Most of the issues we try to solve with technology are actually issues of the organizational systems we live in. And these systems are often defined or at least heavily influenced by political decisions. Think about CovidGraph: The main reason why we need data integration is that data sharing in research and particularly in health care is not incentivized. Physicians still put most of the relevant information in hand written notes, which are not machine readable. Not sharing information can protect your status in a competitive environment. Until you change the rules of that environment.
Why did you join the project - what motivated you and still inspires you today?
I joined the project because I like putting data in graph databases. Even after so many years it’s still fun because it instantly gives you a new perspective of your data. You see it from every angle and you can integrate it with existing knowledge.
I’m still motivated because of our team. It’s incredible that we have this group of people with such diverse backgrounds working on CovidGraph. It’s great to get out of the usual bubble and to work with researchers, large IT companies, freelancers and startups in one team. I think this is special and unusual. Collaboration between industry and academia are encouraged on paper only, at least in Europe. The research funding bureaucracy is actually prohibiting it at every step. I really want to proof that it works and that we can create something that would not be feasible for a research group or a tech startup alone.
Have there been any “uh-huh moments” or surprises during the last months (since the project started)?
Too many to keep track. What really impressed me in the beginning was how fast the visual graph explorer (our pioneer application) was built. Rapid prototyping of end user applications is always important and this shows that a graph database helps to reduce abstraction and makes it easier to build applications. You don’t need ORM or API layers because everyone can understand the data model. I got used to it though and now I expect new features every week 😉
What impressed me most during the project is that so many more people joined and are still contributing today. I look forward to 2021, we have so many great ideas and plans. We should not forget that CovidGraph is open source and does not receive funding. Everyone involved does it because of the team, the challenge and the technology.
What are the challenges of the project for you?
I think it’s important to go beyond the current Covid19 situation and transport our concepts and tools to all areas of health research. The most striking feature of Covid19 is not the incidence or death rate (which we can’t even properly estimate yet) but that it affects many patients on a systemic level. From a clinical perspective it seems to be very different from e.g. Influenza where you don’t see kidney or liver failure. This already points towards the future of CovidGraph: We have to look at the entire system and the interconnection of different diseases.